Some properties and an algorithm of motion planning problem of heterogeneous combinatorial point robots are presented. Heterogeneous combinatorial point robots can be combined and separated freely during moving. It is proven that the problem in a static discrete environment is compliant to the principle of optimality. Dynamic programming algorithms are used to solve this problem. The superposition property of the problem is presented. The time complexity of this problem is O|qV|2kΣ) i=1 nCtn. The motion planning problem of homogeneous combinatorial point robots is a special case of that of heterogeneous ones. A probabilistic roadmap method is used in the experiments and it finds feasible motion plans efficiently.