<p>The contamination of plasticizers in food has attracted more and more attention. In this study, microwave detection technology is used to provide a new method for rapid non-destructive detection of phthalate esters (PAEs) plasticizers in edible oil. Microwave data of rapeseed oil samples containing different concentrations of dibutyl phthalate (DBP) were collected. Two feature optimization methods, competitive adaptive reweighted sampling (CARS) and successive projections algorithm (SPA), were used to select the features of the pre-processed data. Partial least square regression (PLSR) of linear regression model and support vector regression (SVR) and Extreme Gradient Boosting (XGBoost) of nonlinear regression model were established according to the selected best feature subset. Meanwhile, particle swarm optimization (PSO) was used to optimize the parameters of the two nonlinear models. By comparing the prediction performance of the three models, the CARS-PSO-SVR model established after feature selection and parameter optimization achieved the highest prediction accuracy among the evaluated models under the present experimental conditions, and its coefficient of determination (<i>R</i>²) and root mean square error (RMSE) are 0.994 and 0.467&#xa0;mg/kg, respectively. The results show that, using the combination of microwave detection technology and machine learning modeling, high precision and rapid non-destructive detection of DBP content in rapeseed oil can be realized.</p>