In response to the urgent need for consumer behavior analysis and the complexity of massive transaction data, this article conducts in-depth research on user transaction data based on data mining techniques to identify their potential consumption behavior patterns. Firstly, the article conducts data preprocessing operations on the massive transaction data collected; subsequently, key attribute features of the user are generated through feature extraction. Next, the article uses the K-means clustering algorithm to cluster users’ consumption behavior, in order to divide user groups into different consumption types. On this basis, the Apriori algorithm is applied to mine the association rules of user groups and identify the consumption preferences and linkage behaviors that exist between specific groups. At the same time, in order to capture the dynamic characteristics of consumer behavior, the article also uses the ARIMA (Autoregressive Integrated Moving Average) time series model to analyze the cyclical and seasonal trends of consumer behavior. Finally, by constructing a classification model based on random forests, the article predicts future consumer behavior. The experimental results show that in the user clustering experiment, the clustering effect is best when K = 4, with an average contour coefficient of 0.73; in association rule mining experiments, higher support and confidence can effectively improve the accuracy of rules; in the time series prediction experiment, the Mean Squared Error (MSE) of the ARIMA model is 0.8421, and the Mean Absolute Percentage Error (MAPE) is 7.63%. The comprehensive data mining method proposed in this article can effectively identify and predict users’ consumption behavior patterns, significantly improving the accuracy of consumption behavior analysis and proving the advantages and feasibility of the adopted method in analyzing massive transaction data.