In this paper, we investigate the diffusion least mean square (DLMS) algorithm over fading channel, where in addition to channel noise and path-loss the inter-node-interference (INI) among neighboring nodes of a host node is also taken into account. We also analyze the mean-square convergence behavior of DLMS algorithm, under such condition. In addition, based on an upper bound of the derived network MSD, an optimization problem is defined to find an optimal combination strategy. Furthermore, the adaptive version of the proposed combination strategy is presented. Simulation results corroborate the theoretical findings and indicate the superiority of the proposed combination methods over some previously reported algorithms.