Stream temperature is critical for maintaining ecological health and managing water resources. Stream temperature is influenced by a complex interplay of factors, including air temperature, solar radiation, discharge, and land use, with additional complexities introduced by urbanization and climate change. This study explores relationships between land use, riparian vegetation, and stream temperature within the Chesapeake Bay Watershed (CBW), utilizing Long Short-Term Memory (LSTM) model-a type of recurrent neural network model designed for time-series prediction. Here, we use both 30-m National Land Cover Database data as well as new, highresolution (1-m) land cover data developed for the CBW to explore the effects of both reach-scale and upstream watershed-scale riparian land cover on stream temperature. Findings from this work can help us identify stream reaches or watershed areas where land-use change or lack of riparian vegetation have the most significant impact on stream temperature. Such understanding allows us to better prioritize restoration and conservation efforts to mitigate stream temperature increases, thus supporting the development of more targeted environmental management strategies within the Chesapeake Bay Watershed.