This paper presents a comprehensive approach to refine satellite precipitation estimates over the mountainous regions of the United Arab Emirates (UAE) using Long Short-Term Memory (LSTM). The primary aim is to address and correct biases in the CMORPH and IMERG satellite precipitation products by incorporating elevation, minimum temperature, and distance to coast covariates. The study is relying upon the acknowledgement of the substantial inconsistencies that often exist between in-situ gauge data and satellite-derived approximations, particularly in complex terrain areas where orographic effects marginally dominate rainfall trends. LSTM, known for its unique potential to process timeseries data and capture long-term dependencies, is employed to model the spatio-temporal dynamics of precipitation over the UAE's mountainous regions. The study uses daily rainfall data for the duration 2004-2021 in addition to other topographical and climatological variables. The LSTM model was trained with these variables to identify the inherent biases in the CMORPH and IMERG relative to gauge observations. The analysis exhibits that the LSTM-based framework substantially improves the accuracy of precipitation products. The integration of terrain and climatic covariates not only aids in capturing the orographic enhancement of precipitation but also facilitates a more pronounced correction of biases, leading to improved precipitation data quality. The study employs statistical metrics such as the Nash-Sutcliffe Efficiency (NSE), Root Mean Square Error (RMSE), Normalized Mean Absolute Error (NMAE), Probability of Detection (POD), and Critical Success Index (CSI) to quantify the improvements achieved through the LSTM-based bias correction. NSE score was enhanced by almost 45 % for most of the station while the RMSE decreased in the range from 3.4 mm to 0.5 mm. Moreover, corrected products could capture the true rainfall with a 30 % higher accuracy than the raw products as shown by POD and CSI calculations. This study contributes to the field of hydrology and climate science by improving the accuracy of satellite-derived precipitation estimates in mountainous regions and sets a precedent for the application of advanced machine learning techniques in environmental science. The implications of this research are far-reaching, offering valuable insights for water resource management, agricultural planning, and disaster mitigation efforts in the UAE and similar regions worldwide.