<p>While crop yield prediction has mainly focused on cereals like wheat, corn, and rice, developing regression models using field and satellite data for chickpea yield remains underexplored. This study estimated rainfed chickpea yield in Kermanshah County, Iran, by combining field measurements and remote sensing vegetation indices. Data from 14 rain-fed chickpea fields near rain gauge stations were collected during the 2022 agricultural season. The ground-based measurements included soil moisture, soil texture, and leaf area index (LAI). Additionally, remotely sensed indices—namely the Normalized Difference Vegetation Index (NDVI), Soil Adjusted Vegetation Index (SAVI), and Temperature Vegetation Dryness Index (TVDI)—acquired from Landsat 8 and 9 satellites were, incorporated into the regression models. The optimal model was identified by applying the Bayes Information Criterion (BIC), Mallows’ Cp, LASSO, and Elastic Net Regression. This study evaluated model simplicity using R², RMSE, MAE, and Durbin-Watson statistics. The results show that the regression model based on the BIC and Cp criteria during the flowering stage is the most suitable model for estimating rainfed chickpea yield in Kermanshah County. During the calibration phase, the model accounted for 93% of the variation in chickpea yield with an error rate of 8%, identifying Sand and LAI as the primary predictors. In the validation phase, the error rate increased to 11%, while the R² declined to 87%. Therefore, the model can estimate chickpea yield 25 days before harvest with an acceptable error margin relative to the regional average yield (431.4&#xa0;kg ha⁻¹), serving as a criterion for decision-making.</p>