In recent years, students’ comprehensive quality has received more and more attention. Therefore, this paper aims to conduct a comprehensive evaluation of students' performance based on a data analysis framework using the K-means algorithm for clustering analysis. Considering the importance of factors such as students' moral education, intellectual development, classroom performance, and attendance rate in evaluating their overall quality, we selected these factors as features and used the K-means algorithm to group students into different clusters, with each cluster representing a category of students with similar characteristics. We evaluate the overall quality of students by weighting the clustering results with the actual ranking of students' average grades. Before conducting the clustering analysis, we first collected multidimensional data from the students, including academic performance and participation in activities. We then preprocessed the data, such as cleaning and normalizing it, to ensure its accuracy and reliability. Next, we used the K-means algorithm to perform clustering analysis on the processed data and grouped the students into different clusters. For each cluster, we analyzed its characteristics and compared the differences between different clusters. Finally, We weight the clusters with the actual ranking of students' average grades to assess their overall quality. According to multiple experiments, the accuracy of using the K-means algorithm for comprehensive evaluation of student performance is between 70% and 90%, and the efficiency is improved by 50%. The specific numerical effect improvement depends on factors such as the actual dataset and feature selection.