The article introduces a new method for segmenting gray-level images called the Two-Dimensional Otsu's Curve Thresholding Segmentation Method. This method improves upon the traditional two-dimensional Otsu's method by considering edge regions in images, resulting in better segmentation, especially for images with noise. To speed up the segmentation process, a recursive algorithm and wavelet transfer technique are used, reducing the amount of computation needed. The new method only needs to search specific points in the histogram, making the process more efficient compared to the traditional method.