This study has been undertaken to model head rice yield (HRY) and whiteness changes during milling in relation to the physicochemical properties such as amylose content, alkali spreading value, gel consistency and protein content of 10 rice varieties. It was found that the reduction in HRY was a power function of the milling duration with R-2 of fitting ranging from 0.974 to 0.997 and root mean square error (RMSE) values of below 1% among the selected varieties. Also, the parameters of the model describing reduction in HRY during milling were the function of the physicochemical properties of rice. The major contributing physicochemical properties were amylose content (AC) and gel consistency (GC). It was possible to estimate HRY during milling from the physicochemical properties using these functions within +/- 10% of the actual values. The milled rice whiteness could be satisfactorily modeled by a two parameters non-linear model with R-2 of fitting ranging from 0.972 to 0.997 and RMSE values of below 1.0 whiteness index (WI) for all selected varieties. Further, it was found that the parameters of the model representing whiteness during milling were the functions of the physicochemical properties of the rice. The important contributing physicochemical properties were alkali spreading value (AS) and protein content (PC). The whiteness could be estimated from the physicochemical properties using such functions within +/- 5% of the actual values. Thus, it was concluded in the study that HRY and whiteness could be successfully modeled during milling operation in relation to the physicochemical properties of rice. (c) 2007 Elsevier Ltd. All rights reserved.