In this research, the effects of cutting parameters on material removal rate and surface roughness, are investigated. Therefore, after that the comprehensive model of low-immersion milling is developed, the optimum cutting conditions has to be found for optimizing all of them. The stability criterion is considered as the optimization constraint which is calculated by TFEA. On the other hand, instead of using explicit equation for calculating surface roughness, such as previous works, surface roughness is calculated by TFEA for all of the cases that are needed. Finally, the ability of Genetic algorithm, Particle Swarm Optimization and Imperialist Competitive Algorithm for searching optimum cutting parameters are compared and the results are reported. By comparing the results of the three algorithms it is shown that the ICA is more powerful to deal with nonlinearity aspects of the problem and to tackle sticking in local minimums. Also it is demonstrated that the convergence rate of the ICA is faster than the other two methods. Finally, experiments to confirm the changes of the objective function toward optimal point are done and error percentage of objective function at obtained optimal point compared with experimental result is determined.