Features selection methods not only reduce the dimensionality, but also improve significantly the classification results. In this study, the effect of the initialization population using the population factor has been explored. There are twenty wolves obtained by the population initialization method in binary multi-objective grey wolf optimization for features selection. There are two objectives function that will be minimized i.e. number of features and error rate. The proposed method has been compared with the previous study Binary Multi-Objective Grey Wolf Optimization (BMOGWO-S) using UCI datasets, oil and gas datasets. The results reflect that the proposed method outperforms all existence methods in terms of reducing feature numbers and error rates. © 2013 IEEE.