Accurate crop production forecasting is crucial for enhancing food security and guiding agricultural policies in Senegal, a country where agriculture plays a vital role in the economy. Traditional methods, employed by the International Production Assessment Division (IPAD) of the U.S. Department of Agriculture, have shown limitations, particularly in predicting the productions of key crops like rice, corn, millet peanuts, and sorghum. To address these challenges, this study presents a machine learning-based approach integrating climate data rainfall, temperature, and Normalized Difference Vegetation Index (NDVI) to improve forecasts. The methodology applies advanced statistical techniques, including differencing, to remove trends from the data, and various regression models. The work also introduces a feature weighting strategy that accounts for regional differences in crop production volumes, which has proven effective in reducing the complexity of the model while preserving essential information. The results show a significant accuracy improvement, with models achieving a Mean Absolute Percentage Error (MAPE) of 6.4% in 2022 and 7.6% in 2023, cutting errors by 60% compared to IPAD. These findings underscore the potential of machine learning techniques to significantly enhance crop production forecasting in Senegal, providing a robust tool for policymakers to make data-driven decisions regarding agricultural practices, import/export strategies, and pricing policies.