Drug controlled release system is a kind of controlled release system which can achieve specific drug release target by adjusting some design parameters. In order to optimize the initial concentration of drug controlled release system based on time-fractional order diff-usion equation, B-spline wavelet method is used to solve the forward problem. The niche cuckoo search algorithm, which combines the niche strategy with the cuckoo search strategy, is applied to optimize the initial drug concentrations at different fractions, thus approximately achieving the three expected drug release targets. For the solution of the forward problem, an iterative solution scheme of B-spline wavelet method is established by combining the definition of the Caputo derivative and the cubic B-spline scaling function. The problem of optimal design of drug controlled release system is reduced to the problem of parameter identification of fractional diffusion equation based on the inverse problem solution scheme. In order to realize parameter inversion control, the niche cuckoo search optimization algorithm is introduced to inverse the initial concentration in the controlled release system, which effectively solves the problem. The niche cuckoo search algorithm is easily trapped in local extremum. The optimal control parameters are given based on the proposed algorithm for three kinds of release targets: constant velocity release, linear reduced release and nonlinear release. Numerical results show the effectiveness of the proposed method.