This study explores the performance of Temporal Graph Convolutional Networks (TGCN) and Light Graph Convolutional Networks (LightGCN) in recommendation systems, using the MovieLens dataset. Traditional collaborative filtering methods often face challenges such as data sparsity and cold start problems. To address these, graph-based models have emerged, with TGCN incorporating temporal information and LightGCN simplifying the graph convolution process for enhanced efficiency. Our comparative analysis shows that while TGCN excels in capturing temporal dynamics, LightGCN offers significant computational efficiency. These findings highlight the potential of graph-based models to improve recommendation accuracy and user satisfaction.