Wheat is one of the most strategic crops for food security and occupies a large proportion of agricultural lands in Iran. However, climate change, increasing water scarcity, and the limitations of traditional field-based statistics have made accurate monitoring of wheat cultivated areas a challenging task. In this study, an integrated framework based on multi-source remote sensing data and a machine learning approach was developed to identify rainfed and irrigated wheat fields and to estimate their cultivated area in the Maroon watershed in southwest Iran. A set of spectral, radar, and thermal features representing the biophysical, structural, and water-related characteristics of wheat was extracted and used as input variables for the model. The Random Forest algorithm was selected due to its robustness against overfitting and its ability to handle heterogeneous and non-linear data. Model performance was evaluated using a confusion matrix, overall accuracy, and the kappa coefficient. The results demonstrated a high classification performance, with an overall accuracy of approximately 97% and a kappa coefficient of about 0.96 for the test dataset. In addition, the estimated wheat cultivated area showed a low bias of around 3%, indicating a high level of reliability in distinguishing between rainfed and irrigated wheat fields.