An enhanced multiple-edge response (MER) based eye diagram estimation method is proposed to evaluate the performance of nonlinear systems with input jitter. Compared with existing MER-based methods which only took into account the bit effect, the proposed method first determines both orders of bit effect and jitter effect. These decided orders can affirm the necessary MERs. Subsequently, the proposed method figures out the minimal number of sampling points so that the necessary MERs can be recovered quickly based on the Nyquist theory and can be used to create eye diagrams. Lastly, the eye diagrams and their parameters are compared with those generated by traditional transient simulation and an existing MER-based method which introduces input jitter through a convolution process. The result indicates that this enhanced method is more accurate than the existing MER-based method.