Chemical Warfare Agents (CWAs) are toxic chemicals. Among CWAs, nerve agents are lethal compounds at low concentrations, and, thus, there is the need to detect and quantify the danger. The combination of Surface-Enhanced Raman Spectroscopy (SERS) with centrifugal microfluidic (CM) systems is a possibility for a safer handling and a fast analysis of nerve agents.In this work, the SERS signal from single-component solutions and mixtures of Tabun (GA) and VX were analyzed with Au-capped silicon nanopillar (NP) SERS substrates integrated on a CM platform and recorded with a custom-built Raman System (RS). Univariate and multivariate analysis methods were applied for the quantification. Both GA and VX were detected and quantified individually with Partial Least Squares regression (PLSR) models, with LoD and LoQ values of 7.39 ppm and 22.17 ppm respectively for GA, and 7.13 ppm and 21.39 ppm for VX. Detection and quantification of GA and VX in mixtures was achieved with Supported Vector Machine (SVM), obtaining R2 of the prediction 0.87 and 0.95, respectively.The results obtained show the potential of the SERS-integrated CM platform combined with machine learning analysis as a reliable method for CWAs on-site detection and quantification and for national security applications in real-case scenarios.