The brain tumor segmentation acts as a major task in the field of medical image processing. The detection of cancer in manual manner with the large MR images is very tedious, as it requires more time. Thus, in order to overcome the drawbacks of the manual segmentation of the brain tumor, automatic brain tumor segmentation is needed. This paper presents an automatic segmentation of brain tumor using Contourlet transform and Chan-Vese active contour model. The Contourlet transform captures edges and smooth contour at any orientation and it also filters the noise in image in a better way. The Contourlet transform with its extra feature of directionality gives images of high resolution. Chan-Vese active contour models are region based segmentation models and they use optimal piecewise smooth approximation function for the segmentation. These models are very accurate as compared to the traditional segmentation techniques.