The price of rice has a significant impact on the welfare of farmers. The sudden fluctuations in rice prices can significantly impact farmers' decisions regarding the optimal strategy for selling their rice. It is challenging for farmers to accurately predict the price of rice due to the specialized knowledge required to do so. The predominant approach to predicting rice prices is statistical science, which necessitates the involvement of highly skilled experts, extensive preparation periods, and substantial financial resources. Consequently, there is a need for an alternative method to predict rice prices. In recent years, machine learning and deep learning have been widely used to predict rice prices. In this study, we build a rice price prediction model in Bandung Regency using the deep learning method Temporal Fusion Transformer - Grey Wolf Optimizer. The results of the validation process indicate that the Temporal Fusion - Grey Wolf Optimizer model produces the highest Pearson Correlation (CC) value of 0.921 in predicting rice prices in the next five days.