The geometric shape of the human cerebral cortex is characterized by its complex and variable folding patterns. This pattern can be described at different scales from local scale such as curvature to global scale such as gyrification index or spherical wavelet. This paper presents a parametric folding pattern descriptor at the meso-scale of a cortical surface patch. The patch is represented by Bezier Control Points after the Bezier surface parameterization, and the grid coordinates of these points, called BCP image, are used to describe the patch's folding pattern. Based on the intensity pattern of the BCP image, surface patches are classified into different patterns using both model-driven and data-driven clustering approaches. Our experimental results demonstrated that the BCP image is quite effective and efficient in representing the folding pattern of a cortical surface patch.