The efficient, economical, and sustainable production of biodiesel from waste cooking oils (WCOs) depends on the availability of simple, rapid, and low-cost methods to test the quality of potential feedstocks. The aim of this study was to establish the applicability of stochastic modeling of e-nose profiles in the evaluation of recycled WCO characteristics. Olfactory profiles of 10 WCOs were determined using a Sensigent Cyranose® 320 chemical vapor-sensing device with a 32 sensor-array, and a stepwise multiple linear regression (MLR) analysis was performed to select stochastic parameters (explanatory variables) for inclusion in the final predictive models of the physicochemical properties of the WCOs. The most important model parameters for the characterization of WCOs were those relating to the time of inception of the e-nose signal “plateau” and to the concentration of volatile organic compounds (VOCs) in the sensor region. A comparison of acid values, peroxide values, water contents, and kinematic viscosities predicted by the MLR models with those determined by conventional laboratory methods revealed that goodness of fit and predictor accuracy varied from good to excellent, with all metric values >90%. Combining e-nose profiling with stochastic modeling was successful in predicting the physicochemical characteristics of WCOs and could be used to select suitable raw materials for efficient and sustainable biodiesel production.