In view of the existing situation of fruit quality detection in China, which is still dependent on human sense organ to identify and judge the fruit, the broad application prospect of machine vision in quality evaluation of agricultural products, the method to detect the size and surface defect of Huanghua pear by machine vision were studied. The image processing region was decreased greatly by selecting suitable image processing window, and the thinned edge of pear was gained by use of Sober operator and Hilditch thinning algorithm. The maximum diameter, which is perpendicular to the line joined the center of pear and the intersection point of the stem and pear, was adopted to represent the size of pear, and it was found that the correlation coefficient of real size versus detected size was 0.96. In the light of color difference in the joint area of the defected and nondefected area, the light values of R(red) and G(green) were used to find the suspected defected area. The whole defected area was found by means of region growing method, and the area of surface defected was calculated finally. These results laid a solid foundation for further developing fruit quality detection system using machine vision.