In this paper, we provide a homology based algorithm for disaster recovery of wireless networks. We consider a damaged wireless network presenting coverage holes that we need to restore by patching the holes. We propose an algorithm which adds supernumerary vertices to cover the entire area, then run an improved reduction algorithm to reach an optimal result with a minimum number of added vertices. The algorithm is homology based, using simplicial homology representation to compute the coverage and abstract simplicial complex structure to know redundancy vertices information. We present various original methods for the addition of vertices of which we compare the performances between each other and with known methods.