Weapon-Target Assignment problem is the key of air defense command and control. Aimed at the problems that precocity and slow convergence exist in the WTA in air defense, an improved genetic algorithm is proposed. The membership and nonmembership functions are defined for both object and constrain function by importing the theory of intuitional fuzzy sets, and then an intuitional fuzzy WTA model is established on the basis of the “minmax” operator. In the light of constant mutation probability and lack good father individuals in progeny population, a strategy of MetaLamarckian learning for simulated annealing and the adaptive mutation is designed to improve genetic algorithm. The method is effective in solving air defense WTA problem compared with the other algorithms.