Number of experiments has always been a critical criterion in analyzing and characterizing a process. The purpose of this work is to develop a generalized analytical-experimental model based on a typical curve fitting and normalizing method for reducing the number of required experiments to predict expected results. Firstly, the mathematical base of the suggested approach was explained. Then, the feasibility and practicality of the model were evaluated by a case study developed on data of an experimental study in thermal aspects of a creep feed grinding (CFG) process. The models were built by using the input parameters such as: grinding surface temperature and grinding energy at different feed rates, wheel speeds, and cutting depths. Grinding energy and surface temperature generalized models were set up using only the same seven experimental results, and then results of the rest of the experiments were predicted by models. Comparison between the predicted and experimental results shows the capability of the model in prediction of total grinding energy within an average absolute error of 3.8% and grinding surface temperature within an average absolute error of 4.2%. The capability of the approach in decreasing the number of experiments was proved as well. Besides the mentioned advantages, the proposed approach provides models based on both experimental data and mathematical methods which make it free of any assumptions required in mere analytical and numerical models.