In this paper, denoising and segmentation of medical image is performed using morphological filters and watershed algorithm. Watershed Algorithm provides the complete division of image. It has low computational complexity but it suffers from over-segmentation. Segmentation is a process which divides the image into number of segments but it is very sensible to noise. Although technology has been evolved but still, noise may come into image during the acquisition of image either due to instrumental error or environmental factors. So, for obtaining acceptable results of segmentation, it is necessary to eliminate or reduce the amount of noise. For denoising, in this paper various morphological filters are used with the improved watershed segmentation. The proposed algorithm is applied on different medical images like X-Ray, Ultrasound, and MRI and results are evaluated on the basis of MAE, MSE, PSNR and number of segments.