Rainfall causes natural disasters such as floods that people around the world face every year. Rainfall prediction is of great importance for countries like Ethiopia, whose economy is heavily dependent on agriculture. Therefore, its accurate prediction is crucial for the plantation planning and management of farmers. The aim of this study is to investigate the applicability of two machine learning models (i.e., LGBM and RT) for rainfall modeling and compare them with one of the well-known numerical weather forecast models, ECMWF. In this study, we utilized 30 years of rain-gauge data over Bahir Dar station obtained from National Metrology institute, Bahir Dar Sector, from 1993 to 2022. When developing the model, 80% of the data is used for training and 20% for testing. Local climate variables such as relative humidity, dew point temperature, minimum temperature, maximum temperature, wind speed, convective available potential energy, and sunshine as well as large-scale climate variables such as sea surface temperature were used in model development. Four statistical error metrics such as correlation coefficient (R), coefficient of determination (R2 ), mean absolute error (MAE), and root mean square error (RMSE) were used to evaluate model performance. The results of the comparison between the two machine learning models showed that LGBM (R = 0.9046, R2 = 0.8091, RMSE = 4.4066 mm, MAE = 1.7260) outperformed the RT model. For both daily and monthly precipitation variability, the LGBM model produced the highest and best for RT, and the ECMWF produced relatively poor performance. The developed predictive rainfall model for the station produced acceptable performance for daily precipitation with error metrics of R (0.9046), R 2 (0.8091), MAE (1.7260 mm) & RMSE (4.406 mm) and R (0.9813), R2 (0.9575), MAE (1.304 mm) & RMSE (0.7174 mm) for monthly precipitation variability. Keywords: Rainfall modeling, Machine learning (LGBM and RT), Numerical weather prediction, Daily and Monthly rainfall prediction