In this article, an adaptive path following the controller of a multijoint snake robot (MSR) based on the improved Serpenoid curve is proposed. The proposed controller can make the MSR follow the desired path. Compared with the traditional controller, this controller can make the position error possess fast convergence speed and high stability. The swing, the controller can estimate unknown friction coefficients, which improves the adaptive path following the ability of the MSR in an environment with unknown friction coefficients. First, the dynamic model without lateral force is established. Then, the control objectives of the controller are formulated. Third, the Serpenoid gait equation is improved, and the state-dependent time-varying amplitude is obtained. Fourth, the input–output control function of the system and the tracking function of the swing amplitude compensation are designed by the adaptive control method. The stability of the motion attitude angle variable errors and uniformly ultimately bounded stability of the tracking position are verified, respectively. Finally, the effectiveness and superiority of the proposed controller are verified by experiments.