This paper presents the study on the surface roughness model of workpiece in grinding. Based on the analysis of some previous studies, this research identifies the most outstanding model. After analyzing and evaluating its advantages, that model is considered in order to improve. The development of the model is implemented by putting into the roughness model two parameters that significantly affect the surface roughness in grinding, namely the elastic module of grinding wheel and workpiece material. The results of the application of the improved model to the prediction of surface roughness are compared with the roughness values when using CBN grinding wheel to grind C45 steel. It shows that the predictive roughness values are greatly close to the experiment results. The average discrepancy between predictive results with experiment results is approximately 8.11%. This study offers a promising ability to predict the surface roughness of a detail in grinding