In this paper vibration of cutting machine during sawing carbonate and granite ornamental stones was investigated through making a cutting machine on a laboratory scale. For this purpose, properties 7 samples of carbonate stones and 5 samples of granite stones were measured and 211 sawing experiments were performed. Predictives models were developed using different variation of physical and mechanical parameters by incorporating statistical and intelligent methods. The performance of the developed models was evaluated using R2, RMSE, MAPE and VAF criteria for two different types of test datasets consists of Type A and B; data type A included data for rock samples available at the learning stage and data type B included data for rock samples not available in the training phase. The best model for each group of rocks was introduced by taking ranking strategies, evaluation criteria, speed, easiness and reliability of developing method into account. Results indicated that the best model for both rock type was in the form of multivariate nonlinear regression. The similar parameters of these models were depth of cut, feed rate and Schmiazek abrasivity factor. In addition, Young’s modulus and UCS were the special parameters in the carbonate and granite rock models, respectively. These special parameters were in accordance with the finding of sensitivity analysis results.