Image segmentation plays a fundamental role in image processing. Active contour models have been widely used since they handle topological change easily and provide smooth contours. However, noise presents challenges for edge-based level set methods since it leads contours easily passing through objects or falling into local minima. In this paper, we propose a weighted edge-based level set method based on multi-local statistical information to better segment noisy images. Through analysing the deficiencies of constant length and regional coefficients and traditional edge stop function in noisy image segmentation, weighted length and regional coefficients and modified edge stop function are proposed to overcome their shortcomings, respectively. The weighted edge-based level set method is used to segment synthetic and real images that have added different types and levels of noise. The experiments indicate that our method provides higher segmentation accuracies and more accurate segmentation results, which demonstrate its effectiveness and robustness.