Abstract: This paper establishes the mathematical model for the nonlinear optimization of complex mine ventilation network. The objective function is the fewest regulation facilities and the smallest air power of the main fans; the decision variables are composed of the air quantity of the cotree branches which are not the constant branch, and the resistance of the regulation branch of the maximum spanning tree. Based on the standard DE algorithm, and using the greedy searching strategy and parent individual neighborhood depth searching method, the offspring is generated by the best parent individuals, then the modified differential evolution algorithm is put forward to solve the nonlinear optimization of mine ventilation network. The practical application shows that: the modified differential evolution algorithm is feasible, and it obtains an air regulation plan which ensures mine ventilation system operating safely, economically, as well as reliably.