The traditional methods for determining pesticide concentrations are time-consuming, complicated, and require extensive pretreatment processes. In this study, near-infrared (NIR) spectroscopy was used to determine trace chemicals. The dry-extract system for infrared (DESIR) technique was used to prepare samples. Filter paper was used as the substrate. Pesticide solutions were prepared by dissolving a commercial pesticide in distilled water at different concentrations (1.25 to 400 mg kg-1). Samples were prepared by pipetting the solution onto the filter paper and then evaporating it in a vacuum drying oven. Spectral curves of the samples were acquired in the range of 10000 to 4000 cm-1 using an NIR spectrometer. Partial least squares regression (PLSR) was used to establish prediction models. The best prediction result was obtained using PLSR with multiplicative scatter correction (MSC) and first derivation as the pretreatment procedure. The process was able to predict the concentrations of chlorpyrifos with R = 0.899. A support vector machine (SVM) was used to establish a classification model. The result showed that 89.286% of samples were correctly predicted when the sample set was divided into three classes of chlorpyrifos content ( 300 mg kg-1).