Compared with wireless sensor networks and mobile ad hoc networks, wireless mesh networks mainly focus on improving the throughput of multicast, while interference severely limits the network throughput. When building a multicast topology, the minimum cost or shortest path is generally taken into account in the traditional methods, and only a few works have tried to improve the performance by reducing interference. However, they calculate the interference by the method for unicast topology, which is not suitable for multicast. For example if n nodes will receive simultaneously packets from one node, among these n nodes there is interference in unicast, but not in multicast. Therefore this tudy proposes the multicast conflict graph to calculate interference of the multicast topology, and then the concise definition of interference of multicast trees is provided. The study shows that building minimum interference multicast trees (MITs) is a NP-complete problem and proposes a gravitation-based heuristics to approximate such optimal trees. To apply to the environment of multi-channel, the study also proposes the multi-hop channel algorithm (MH) for multicast, which can meet different interference ranges. Simulation results reveal that the algorithms can reduce interference and increase throughput in both single-interface single-channel and multi-interface multi-channel wireless mesh networks.