Wireless body area networks (WBANs) are communication networks of sensors (and/or actuators), which can be placed on, inside, or around the body and represent a new generation of personal area networks. In contrast to the typical wireless sensor networks (WSNs), WBAN sensors are tiny and have limited power supply. Due to irregular human body movement and WBAN sensor distribution, fading can cause a lot of error, which can be determined by measuring the bit error rate (BER). A higher BER requires more power. Therefore, to reduce BER, power consumption (PC) and duty cycle (DC), a game theory-two master nodes-based cooperative protocol (GT-TMNCP) is proposed in this work. DC is enhanced by a further 13% using the proposed GT-TMNCP, leading to a BER reduction 3 times greater when compared to TMNCP, and PC improved by 14%, lower than with the TMNCP, for shadowing at 5 dB, critical index data (IC), IC = 3 and 5, representing maximum critical data. The results also demonstrate that BER is reduced up to 8 times, DC is enhanced by 8%, and PC by 7% for shadowing 9 dB, IC = 7, compared to the QoS optimisation approach and TMNCP.