This paper explores an intelligent self-learning PID control strategy for uncertain disturbed electro-hydraulic servo system with parameter uncertainty and external load disturbance. This method is a data-driven control strategy which is independent of the precise model of the system. Firstly, the linear affine electro-hydraulic servo system with the time-varying parameter term and nonlinear uncertainty term is devised using an improved dynamic linearization method. Then, the gradient estimation method and the time difference method are utilized to estimate the time-varying parameter term and nonlinear uncertainty term respectively. Furthermore, the lost information of over-linearization is compensated by additional error information from iPID. Finally, the parameter updating law and learning control law of the electro-hydraulic servo system are designed according to optimal criteria. The convergence of control strategy is proved by theoretical analysis and simulation experiments. And the superiority and accuracy of the method are verified by comparative experiments. The paper shows that the adverse effect of nonlinear disturbance is suppressed and the accurate tracking of ideal trajectory is realized.