Abstract: We propose a PID (proportion-integration-differentiation) parameter optimal iterative learning control algorithm based on singular value decomposition to solve the tracking control problems of discrete linear systems. The traditional parameter optimal iterative learning control algorithm can guarantee tracking errors converging to zero only under the condition that the original plant is positive-defined. In order to overcome this limitation, the proposed algorithm establishes the norm optimal performance index and obtains the learning gain matrix by applying singular value decomposition to the original plant. The algorithm guarantees that the closed-loop tracking errors of this algorithm converge monotonously to zero even when the original plant is non-positive. We apply a PID controller to the design of parameter optimal iterative learning control to improve the learning efficiency of this algorithm. Theoretical analysis and relevant proof of the convergence properties of this algorithm are also given. The result of the simulation verifies the effectiveness of the proposed algorithm.