Based on the principle of clonal selection, an improved adaptive immune clonal selection algorithm is put forwarded by using real-coded method, where adaptive mutation operator is used according to the changes of function value and mutation coefficient decays with the iteration. In addition, the strategy of mutation by section is used according to the iteration. To test its efficiency, optimization computing is conducted for five typical test functions. Subsequently, a complex nonlinear optimization model is solved based on the algorithm. The results show the validity of this algorithm to solve the above problems.