In order to enhance the current tracking control performance of the active power filter, this paper proposes an intelligent terminal sliding mode control scheme based on a continuous radial basis emotional neural network. Firstly, the mathematical model of an active power filter including lumped uncertainty is introduced. Then a recursive terminal sliding surface composed of a fast non-singular terminal sliding surface and a recursive terminal integral sliding surface is constructed, which can not only ensure that the tracking error converges to zero in a finite time, but also eliminate the arrival stage of the sliding surface by setting appropriate initial values for the sliding surface parameters. In addition, in order to effectively deal with uncertainty, a continuous radial basis emotional neural network is utilized to approximate the uncertain parameters of the system, and its stability and convergence is also ensured using the Lyapunov method. The designed continuous radial basis emotional neural network is not only simple in structure, fast in response, but also has the ability to adjust parameters online. Simulation and experimental results show that the control scheme has excellent current tracking ability and anti-interference ability.