To improve the sales volume of electronic commerce platforms, predicting user purchasing behavior has become an important research area. This paper addresses the issue of user purchase prediction in e-commerce platforms. Firstly, a machine learning-based user purchase prediction model is proposed. Secondly, a machine learning decision model based on random forest algorithm is established, which extracts features from various aspects such as user historical purchase data to determine customer purchasing behavior. Finally, the trained machine learning model is used to predict customer purchasing behavior. Experimental results show that the proposed model exhibits excellent performance in terms of prediction accuracy, fitness, and has practicality and promotion value.