To evaluate the software behavior of the electronic control unit (ECU) of automotive electrical parking brake (EPB), a software-in-the-loop (SiL) simulation system is built. The EPB is simulated by ARX (auto-regressive with auxiliary input) model, ARMAX (auto-regressive moving average with auxiliary input) model, and NNARMAX (neural network ARMAX) model. By system identification, the ARX(3,4,2), ARX(4,4,2), ARMAX(3,3,1,1), and ARMAX(4,4,3,2) models are derived. Validation results show that the four-order ARMAX model and the NNARMAX model better simulate the actuator of the EPB.