Soft Magnetic Composites (SMCs) have huge application potential in electrical devices. With the development of additive manufacturing (AM), the properties of SMCs can be boosted. It is therefore required an optimization algorithm to design SMCs considering the AM capacity. Since the magnetic and loss properties of SMCs are highly sensitive to the shape evolution of the components, we introduce a topology optimization approach with integer programming. At each iteration, the components of the composite are determined with crisp boundary. This optimization algorithm converges rapidly because of the topology gradient and the optimal topology is reached.