In a smart grid environment, wireless sensor networks (WSNs) perform a significant role in supervising and taking control decisions based on the data transferred from source to destination. The work in this paper is aimed to develop a cross-layer multi-channel MAC (CMCMAC) protocol for interference mitigation between ZigBee and WiFi operating in common 2.4GHz frequency band. In this protocol, to avoid the interference between the channels, the future interference is predicted using the Hidden Markov Model (HMM), based on the prediction a channel with the minimum interference is selected for data transmission to enhance the network performance. The results obtained show the enriched network performance. © 2016 IEEE.