Modelling is a common subject in terms of application areas throughout many scientific fields such as statistics, mathematic, machine learning, artificial intelligence, deep learning, engineering, bioinformatic and so on. This interdisciplinary attribute nudge modelling into creating a simple representability of its results. The process is made by performance measures to show how well or worse the results of the model. Performance measures can be a single scalar, ratio, vector, diagram or even much complex indicators. Performances are actually models' identities and measures are the identifiers of these models. Therefore, measuring the performance of a model is very crucial step and choosing the most proper measure directly effects the identification of related model. Performance measuring can also be used for the selection of the best model among them. There are various performance criteria in the literature to determine the best model. On the other hand, it is very difficult to say which performance criterion is better according to the analysis method used or the data analyzed. Therefore, how to determine the best model is a very challenging task. In this study, to deal with this challenging issue, a new intelligent model evaluation method is proposed to measure the performance of models by approaching human intelligence. Therefore, the proposed intelligent approach offers a new perspective in model evaluation.