In the context of Computer-Supported Collaborative Learning (CSCL), it is essential to propose activities tailored to students’ profiles and interactions. This study explores the use of a recommendation system based on the K-Nearest Neighbors (KNN) algorithm to suggest relevant collaborative activities to learners. Our approach analyzes student characteristics (learning level, forum participation, activity preferences, engagement) to identify similar profiles and recommend appropriate activities. After normalizing and structuring this data, we apply KNN to determine the K most similar students and suggest activities based on their past experiences. An experiment was conducted with 60 Master's students, divided into two groups: one receiving personalized recommendations and a control group without recommendations. The results indicate a 25% increase in participation rates for the group benefiting from recommendations. Additionally, a t-test analysis revealed a statistically significant difference (p < 0.05) between the performance of the two groups, with an average 15% improvement in scores for students receiving recommendations. These findings suggest that using KNN for collaborative activity recommendations enhances the CSCL learning experience by improving student engagement and performance.