Abstract: In this paper, a principle prototype based on the dielectric property for automatically inspecting the internal quality of apple is introduced. The dielectric parameters of inspected apple are sampled by means of two capacitive sensors and a Inductance-Capacitance-Resistance(LCR) instrument connected with the computer in the principle prototype. The equivalent capacitance of inspected apple, one of dielectric parameters, is used as characteristic for sorting. To a batch of apples, the optimal threshold for sorting can be obtained by on-line studying from some samples of high quality and bad quality apples. The study result indicates that the principle prototype can better classify the apple with different qualities. When the diameter of the inspected apple is between 7.0 cm to 8.5 cm and the diameter of defective part of a bad quality apple is larger than 2 cm, according to the optimal threshold obtained by studying, the accuracy for classifying is larger than 80% with 1 kHz working frequency of capacitive sensor, and its work-efficiency can attain one apple per 2.5 s being close to the efficiency of manual sorting.