Analytical procedures typically consist of a number of equally important steps for sampling, sample treatment, isolation of the target compounds, identification, quantification and data handling. All operations and manipulations carried out with samples before instrumental determination of the tested compounds are considered to be part of the sample treatment/preparation step. Sample preparation would consequently include from labelling and mechanical processing and homogenisation of the studied matrix, to any type of gravimetric or volumetric determination carried out to characterise the analysed (sub-)sample, as well as all subsequent treatments designed to decompose the matrix structure, to perform the fractionation, isolation and enrichment of the target analytes from any potential interference, to make the tested compound(s) compatible with the detector (e.g., phase exchange and derivatization reactions), and to improve their detectability. Nevertheless, the term sample preparation has typically been associated to the latter group of chemical operations, all earlier mechanical and basic treatments being named as sample pre-treatment [1]. This will also be the terminology applied in the present review article. Considering the nature and goal of most sample preparation operations, it is evident that this part of the analytical process has a profound influence on both the total time required to complete the analysis and the quality of the results obtained. However, it has only been in recent years that this step has risen to the prominent place that it now holds within the analytical protocol. The development of trace-level determinations in environmental and food samples have been identified as generating the stimulus for much of the progress in this research area [2]. Whatever the original incentive, it is clear that the continuous demand for accurate and faster determinations of a constantly increasing number of analytes at decreasing concentrations in these complex matrices, together with the increasing interest for the analysis of biological samples and the development of the –omics sciences, have spurred investigations in this active research field. Despite the many efforts carried out during the last two to three decades to improve the techniques used for sample preparation, the sample treatment procedures in use in many application areas are still tedious multistep protocols involving repeated manual manipulation of the extracts. Because of the frequently low concentrations at which the target analytes should be determined, the first step of these protocols usually consists of the exhaustive extraction of the analytes from the matrix in which they are entrapped. The essentially non-selective nature of this initial step makes subsequent purification of the obtained extracts before final instrumental determination mandatory, unless (separation-plus)- detection is highly selective. The several analytical treatments involved in these purification protocols are usually carried out offline, which significantly affects throughput and analysis cost both in terms of time and reagent consumption, makes the procedures prone to contamination and degradation of the analytes, and often results in the generation of relatively large amounts of waste. These features explain why sample preparation is estimated to accounts for two-thirds of the total analysis time and, more importantly, is considered to be the primary source of errors and discrepancies between laboratories [3]. In other words, proper selection and optimisation of the sample preparation scheme are key aspects within the analytical process that can greatly affect the reliability and accuracy of the final results [4,5]. Conventional techniques, such as liquid–liquid extraction (LLE), solid–liquid extraction (SLE) and Soxhlet extraction, are still widely accepted and used for routine applications and/or for reference purposes. However, in recent years, some of these techniques have been revisited and upgraded versions, in which their most pressing shortcomings have been solved, are now available. The studies in this field have also led to the development of new faster and more powerful and/or versatile extraction and preconcentration techniques [6]. Thereby, in many instance, partial and even full hyphenation and automation of the analytical process, or at least of the several treatment steps, are now possible. In addition, sample preparation approaches that fulfil the goals of green analytical chemistry [7] are also available. For obvious reasons, the ideal situation would be the complete elimination of the sample preparation step from the analytical process. However, despite the current degree of development of the analytical instrumentation used for final determination in most instance this is not feasible. Concepts like miniaturisation, integration and simplification became key concepts that have already been proved to effectively contribute to solve some of the drawbacks of conventional sample preparation methods and that, in some studies involving size-limited samples, can probably be considered the best, if not the only, analytical alternatives. The present review article focuses on sample preparation, with examples primarily related to liquid and solid matrices, and more specifically on selected modern techniques, i.e. those introduced in the last two decades or so. Most recent developments and achievements in the field will be discussed on the basis of representative examples. Attention will focus in the analysis of trace organic compounds due to the difficulty associated to this type of determination. Nevertheless, if relevant, examples will also be taken from other application areas as far as they involved a chromatographic (or closely related separation) step for the final instrumental determination of the target compounds. –Omic sciences remain out o