This paper introduced a method of recognizing the partially occluded eggplant based on generalized Hough transform in three-dimensional space. The shape of eggplant was approximately described by using generalized-cylinders from the view of biology. Different edges of projection were obtained through coordinate transformation and 4-D parameter index tables were established. Generalized Hough Transform was adopted to count the potential pose of object and the real pose was screened by comparing the area-ratio of different rotation angles. The experiment demonstrates that the relative errors of position and pose detecting for target are lower than 1.3% and 5.5% by using Generalized Hough Transform, and this method is feasible and effective to recognize different poses and partially occluded objects