In this paper, we propose an algorithm based on basic simulated annealing (SA) to optimize unsteady chemical systems. By parametrizing the control inputs, the system described by a set of differential-algebraic equations is first transferred into a nonlinear programming (NLP) problem. By using a linear function, we furthermore convert discretized control inputs and time grids into the control profile with a variable time length to improve numerical quality. Thus, the discretized control inputs and the corresponding execution time lengths will he considered as a set of decision variables. Then, these decision variables are globally determined by our algorithm with the help of a special integrator to optimize the performance index. In order to exhibit the facility of the proposed algorithm, several typical examples are provided.