In order to realize the automated estimation of orchard yield, we studied on the recognition method of green apple based on machine vision. An image capturing system composed of a color camera and an active lighting system was used to capture images of apple trees at night. A hybrid classifier including an SVM method based on the advantage of H, S and normalized g and a Super-G method was developed to remove the background. The connected apples were found by the area feature and the ratio of major axis to minor axis of the individual regions. Then we used the Euclidean distance transformation method to generate the distance image of the founded connected apple regions. Finally a watershed method was used to separate the connected apples, and the total number of the apples in image were known. The analysis of experimental results regarding 64 images showed that the average rate of correct recognition was 89.30%.