<p>Next Basket Recommendation System aims to predict the series of items a user is likely to purchase next based on their historical interaction data. While existing next basket recommendation models focus on user-item relationships and latent temporal patterns, they often neglect the impact of actual temporal information in interaction data. To address this issue, this study proposes a novel framework that explores how multi-perspective temporal information fusion can enhance prediction accuracy. The framework consists of two key modules: the latent temporal information extraction module, which analyzes the item perspective to capture the evolution of user preferences; and the actual temporal information extraction module, which examines the time perspective to reveal the preference differences of users at different time points. By fusing multi-perspective temporal information, the framework can more comprehensively capture the dynamic changes in user interests, thus improving the accuracy of recommendations. Extensive experiments on four real-world datasets validate the effectiveness of the proposed framework and emphasize the importance of temporal information in next basket recommendation.</p>