This study addresses the satellite swarm strength allocation strategies generation problem with complex constraints. The satellite swarm confrontation attrition process is described by an improved Lanchester equation and multi-phase confrontation patterns. On the basis of them, a satellite swarm strength allocation strategies generation optimal control model is first presented. Unlike the general optimization model, this model considers three constraints such as sunlight angle, supplement number, and confrontation time to make it suitable for describing the space confrontation process. Then, the multi-interval orthogonal collocation (MIOC) method is proposed to generate the strength allocation strategies from the constructed model. This method not only solves optimal control problems but also solves discontinuities caused by multi-phase confrontation patterns. Finally, a commencement solution exploration approach is proposed to explore the appropriate initial value for the MIOC method. The aim of it is to avoid initial value-sensitive problems and to make convergence fast. The effectiveness of the MIOC method is verified through three numerical examples in comparison with the static optimization methods.