Rice is an essential cereal consumed by around 2.5 billion people worldwide, and Brazil stands out among the top ten producers. Brazilian production is recognized for its productivity, technology, and monitoring, mainly concentrated in Rio Grande do Sul, contributing to about 70% of the total production. Like any agricultural commodity, the price of rice is subject to market forces, influenced by various factors such as weather conditions, input prices, and demand, reflected by the population's purchasing power. Price fluctuations can be detrimental to both consumers and producers, especially considering the 5-month period between planting and harvesting. Given these concerns, the main objective of this work is to develop machine learning models capable of predicting the price of this commodity, considering a 5-month horizon and using variables representing supply and demand. While there is existing research aiming to predict the price of rice and other agricultural commodities using different machine learning models, no studies were found specifically addressing forecasting with the same lead time as this work, nor using variables representing supply and demand. Therefore, this project fills this gap. For this research, various machine learning models were adopted, applied both with and without the Recursive Feature Elimination (RFE) technique, using subsets of training and test data with different periods. Additionally, two data adjustment procedures were performed to forecast 5 months in advance: one through direct lagging and another using simulated independent variables, as explained in the Materials and Methods chapter. The results revealed that it was possible to develop such models, which had an average error of approximately 17%, with higher errors noted in specific periods, especially in the second half of 2020. The best-performing model in the 5-month-ahead prediction was the Extreme Gradient Boosting with RFE technique in the direct lagging procedure, achieving a MAPE of 10%.