Focusing on the issue that conflict is hard to detect accurately and comprehensively in collaborative design, a conflict detection model based on constraint was proposed. Considering the hierarchical constraints and constraint satisfaction, the detection model divided constraints into two sets: one set is with known constraints and the other set is with unknown constraints. The constraints of two sets were detected respectively. The set with known constraints was detected by interval propagation algorithm. Meanwhile, Back Propagation (BP) neural network was used to detect the set with unknown constraints. Immune Algorithm (IA) was utilized to optimize the weights and thresholds of BP neural network, and the steps of optimization process were put forward. In the comparison experiments with BP neural network optimized by Genetic Algorithm (GA), the convergent speed was increased by 69.96%, which indicated that BP neural network optimized by IA has better performance in convergent speed and global searching ability. The constraints were described by eXtensible Markup Language (XML), so that computers could automatically recognize and establish the constraint network. The implementation of conflict detection system based on constraint satisfaction was designed. Taking co-design of wind planetary gear train as an example, a conflict detection system in collaborative design was developed on Matlab with C#. The conflict detection model is proved to be feasible and effective, and provides a solution of conflict detection for collaborative design.