The increased growth of e-commerce has generated a vast amount of data that reflects customers’ behavior. This special data creates valuable information that provides an opportunity to reach intelligent retail analytics. This study presents a new Machine Learning (ML) based framework for analyzing customer behavior and predicting product priority in online retail. Utilizing a real-world dataset (Online Retail dataset), the study proposes rich preprocessing, feature engineering, and data transformation to construct new attributes related to products. A classification model is then created using the newly generated attributes to group products based on consumer behavior and purchase trends. Cross-Industry Standard Process for Data Mining (CRISP-DM) is the methodology used to accomplish this work. Three attractive supervised ML algorithms named K-Nearest Neighbors (KNN), Linear Discriminant Analysis (LDA), and Support Vector Machine (SVM) were used to train and assess the clean version of the dataset. Trained models were evaluated using accuracy, precision, recall, and f1-score metrics along with the ROC visualize metric. Moreover, the study provides additional analytic information like feature ranking, correlation analysis, and visual trend exploration to provide deep insight into the underlying customer behavior. Experimental results show that KNN gave the best performance of 98.36% outperforming other models. In addition, this study enhances the prediction of customer behavior, it also helps retailers to gain comprehension of dynamic shopping trends and eventually facilitates data-driven decision-making in online competition.