This paper introduces a methodology for obtaining accurate continuous-time multiple-input, multiple-output models of nonlinear dynamic systems with Wiener characteristics. The models are obtained from complete reliance on experimental data, and this work demonstrates the effectiveness of optimal statistical design of experiments (SDOE) to fully obtain Wiener models. This method is evaluated on a highly nonlinear continuous stirred tank reactor, and its performance is compared to conventional discrete-time Wiener modeling (DTM) using a pseudo-random sequence design (PRSD) and the same SDOE as the proposed method. The proposed method greatly outperforms the DTM developed from PRSD and moderately outperforms the DTM based on SDOE.