 This research explored the potential of multimodal deep learning for improving the accuracy of rice yield predictions using UAV-based multispectral imagery and weather data. The best models were those trained with weekly weather data, which resulted in higher prediction accuracy than other models. Additionally, the spatial patterns of the predicted yield maps differed among models, although the overall prediction accuracy was similar. This suggests that further assessment of the robustness of within-field yield predictions should be conducted before implementation of precision agricultural technologies. This article was authored by M.D. Surajmiya, Rayoya Taunabey, Luth Van Nir Habibi, and others.