<p>The electro-mechanical brake (EMB) system streamlines the structure of automotive braking systems, enhances braking precision and response speed, and significantly improves overall braking performance. However, the EMB is characterized by random uncertainty in structural parameters and exhibits strong nonlinear behavior. Variations in these structural parameters can cause significant changes in system performance, making deterministic studies insufficient to accurately capture its behavior under real-world operating conditions. Therefore, investigating the impact of random uncertainty in structural parameters on EMB performance is of substantial practical importance. To address this challenge, this paper employs the polynomial chaos expansion (PCE) method to systematically model and analyze stochastic uncertainty. Three critical structural parameters—motor rotational inertia, ball screw lead, and brake clearance—were selected as random variables, and their influence on system performance was evaluated. A sensitivity analysis was conducted to quantify the effects of these parameters on system behavior. Subsequently, the weighted standard deviation of braking clamping force and motor rotate speed was adopted as the optimization objective. A multi-objective optimization process was carried out to refine three structural parameters. The optimization results reveal significant improvements in both the weighted standard deviation of the braking clamping force and motor rotate speed, demonstrating an overall enhancement in system performance. To further validate the simulation and optimization results, a bench test of EMB systems was conducted. The results indicate that the weighted standard deviations of braking clamping force and motor rotate speed reduce 4.48 % and 15.25 %, respectively. This study confirms the feasibility and effectiveness of the PCE method in addressing stochastic uncertainty problems in complex nonlinear EMB systems.</p>