In this paper, the starting point generation strategy for parametric optimization problem is promoted to solve complex parametric dynamic optimization problems (PDOPs), and an efficient algorithm framework is developed. Since the starting point strategy is designed for nonlinear programming problems, the PDOPs are discretized by IRK method at first. Then, several multivariate scattered data fitting methods are used to generate the advanced starting points (ASPs) for the discretized models. According to the existence and uniqueness of the solutions of differential equations, a partial ASP strategy is proposed. The novel strategy greatly compresses the empirical data storage and guarantees the solving efficiency simultane-ously.