An acetonitrile-based extraction method for the analysis of 169 pesticides in soya grain, using liquid chromatography-tandem mass spectrometry (LC-MS/MS) in the positive and negative electrospray ionization (ESI) mode, has been optimized and validated. This method has been compared with our earlier published acetone-based extraction method, as part of a comprehensive study of both extraction methods, in combination with various gas chromatography-(tandem) mass spectrometry [GC-MS(/MS)] and LC-MS/MS techniques, using different detection modes. Linearity of calibration curves, instrument limits of detection (LODs) and matrix effects were evaluated by preparing standards in solvent and in the two soya matrix extracts from acetone and acetonitrile extractions, at seven levels, with six replicate injections per level. Limits of detection were calculated based on practically realized repeatability relative standard deviations (RSDs), rather than based on (extrapolated) signal/noise ratios. Accuracies (as % recoveries), precision (as repeatability of recovery experiments) and method limits of quantification (LOQs) were compared. The acetonitrile method consists of the extraction of a 2-g sample with 20 mL of acetonitrile (containing 1% acetic acid), followed by a partitioning step with magnesium sulphate and a subsequent buffering step with sodium acetate. After mixing an aliquot with methanol, the extract can be injected directly into the LC-MS/MS system, without any cleanup. Cleanup hardly improved selectivity and appeared to have minor changes of the matrix effect, as was earlier noticed for the acetone method. Good linearity of the calibration curves was obtained over the range from 0.1 or 0.25 to 10 ng mL(-1), with r(2)>or=0.99. Instrument LOD values generally varied from 0.1 to 0.25 ng mL(-1), for both methods. Matrix effects were not significant or negligible for nearly all pesticides. Recoveries were in the range 70-120%, with RSD<or=20%. If not, they were still mostly in the 50-70% range, with good precision (RSD<or=20%). The method LOQ values were most often 10 microg kg(-1) for almost all pesticides, with good repeatability RSDs. Apart from some minor pros and cons, both compared methods are fast, efficient and robust, with good method performances. The two methods were applied successfully in a routine analysis environment, during surveys in 2007 and 2008.