This paper examines the use of hybrid machine learning models in forecasting customer’s purchasing behavior based on e-commerce data. With the growth of online sales, it became a requirement for businesses online to anticipate customers’ tastes and buying behavior in order to maximize marketing efficacy and customer satisfaction. The study combines traditional machine learning models like Random Forest (RF), XGBoost, and Support Vector Machines (SVM) with Artificial Neural Networks (ANN) in a bid to create robust models with the ability to detect complex patterns in the data. The preprocessed consumer browsing, and transaction data is preprocessed to handle missing values, normalize features, and encode categorical variables. Several performance measures such as accuracy, precision, recall, F1 score, and AUC-ROC are used to measure the performance of the hybrid models. Of the results, the XGBoost with ANN hybrid model achieves the highest performance at 97% at epoch 30, followed by RF with ANN at 95% and SVM with ANN at 96%. These findings illustrate the strength of using hybrid methods in improving predictive accuracy, where XGBoost is more precise and converges at a quicker pace. The research provides significant insights on how to utilize machine learning in e-commerce analysis and consumer behavior prediction, showing the strengths of combining old algorithms and neural networks. The results demonstrate that hybrid models can also be successfully used to present the same predictive tasks in other industries with an even more precise and flexible solution to capture consumer interests in dynamic market environments.