This paper introduces a holistic approach to predict short-term crop yields utilizing a combination of K-Nearest Neighbor (KNN) Algorithm and Random Forest models. The methodology initiates with preprocessing the raw agricultural data, followed by the standardization of relevant crop attributes. This normalization addresses discrepancies in data magnitudes, subsequently elevating prediction accuracy. The KNN nonparametric regression model is then implemented using the Python programming language, forming the core of the short-term crop prediction system. To validate the proposed methodology, actual crop yield data from a selected agricultural region is employed. A comparative analysis explores both average and weighted KNN nonparametric regression models.