A network time-delay compensation method based on time-delay prediction and implicit proportional-integral-based generalized predictive controller(PIGPC) is proposed. The least squares support vector machine(LSSVM) is used to predict the current time-delay, the parameters of the least squares support vector machine are optimized by particle swarm optimization(PSO) algorithm,and the predicted time-delay is used instead of the actual time-delay as the parameters of the network time-delay compensation controller. In order to improve the compensation effect of implicit generalized predictive controller(GPC), this paper puts forward an implicit generalized predictive control algorithm with proportional-integral-based(PI) structure and designs the controller based on implicit PIGPC. Through the simulation results, the effectiveness of this design in the paper is verified. A network time-delay compensation method based on time-delay prediction and implicit proportional-integral-based generalized predictive controller (PIGPC) is proposed. The least squares support vector machine (LSSVM) is used to predict the current time-delay, the parameters of the least squares support vector machine are optimized by particle swarm optimization (PSO) algorithm, and the predicted time-delay is used instead of the actual time-delay as the parameters of the network time-delay compensation controller. In order to improve the compensation effect of implicit generalized predictive controller (GPC), this paper puts forward an implicit generalized predictive control algorithm with proportional-integral-based (PI) structure and designs the controller based on implicit PIGPC. Through the simulation results, the effectiveness of this design in the paper is verified.