This paper presents a prediction method for the laser cleaning thickness of composite coatings using a support vector regression (SVR) model optimized by a particle swarm optimization (PSO) algorithm. Laser cleaning experiments were conducted on aluminum alloy substrates coated with a 20μm green epoxy primer and a 40μm white polyurethane topcoat. An SVR model was developed to establish the correlation between process parameters and the thickness of paint removal for composite coatings. The PSO algorithm was employed to optimize the SVR model′s penalty coefficient C and kernel function parameter g. The experimental results show that compared with SVR model and BP neural network model, the model is more accurate in predicting the laser cleaning thickness of composite coating. The coefficient of determination of the model is 0.96171, the root mean square error is 1.738, and the average absolute error is 1.516 2. In this study, a prediction model of paint removal thickness with high precision can be obtained, which can effectively predict the thickness of laser paint removal, and lays a foundation for further research on intelligent control of composite paint layer.