The higher education system in India is curious about the success of students in education during their study. These educational institutions are adopting several methodologies to improve the quality of education and to improve the success rate of students each year. This is used as a primary objective in improving the academic excellence of a student in his/her higher educational level. The main aim of the study is to create a model that classifies the instances correctly to predict the performance of students using PAFT methodology. The PAFT methodology consists of several attribute of a student modeled into a three-tier model that is collected based on several level testing done on a particular student. The three tier model involves his complete academic details, his/her creative and other interpersonal skills and finally the level of interest towards the present educational approach. To classify the instances correctly, Genetic algorithm is modified in its mutation level with NAND gate. The proposed classifier eliminates the regression fit problem during the selection stage with the help of Tobit regression evaluation of each individual. The classifier is also compared with other techniques like genetic-OR, genetic-AND, Multi-Layer Preceptor and artificial neural network. This classifier optimizes correctly the attributes given as an input for its processing and better learning. The PAFT methodology combined with a genetic - NAND algorithm proves successful in terms of its classification rate. This could be inferred finally that this could be utilized in institutions in determining the performance and success ratio of students as a part of their knowledge management system. Also, this model could also be used for predicting the improvement level of students for the fore coming students based on the collected data set. © 2016, Institute of Integrative Omics and Applied Biotechnology. All rights reserved.