Motion planning in autonomous driving requires accurate perception of the surroundings and precise vehicle localization. However, such accuracy is often unavailable due to inherent localization and perception uncertainties. In this paper, we propose an innovative uncertainty-aware motion planning method. By incorporating uncertainty constraints into the Constrained Iterative Linear Quadratic Regulator (CILQR) framework, our approach significantly improves planning efficiency. Specifically, an occupancy probability map-based uncertainty constraint is seamlessly integrated into the CILQR framework to enhance the planning system’s flexibility and adaptability. To further address the combined effects of localization and perception uncertainties, an extended-ellipse-based uncertainty constrain is designed for straightforward integration into the CILQR framework. Extensive simulations conducted in the CARLA environment demonstrate that our method outperforms existing approaches in terms of both success rate and computational efficiency.