In order to reduce the disadvantages of current biological product quality testing methods, taking the quality testing in cocoon trade markets as an example, this paper has proposed a quality nondestructive testing method for biological products based on the analysis of vibration signal from acceleration sensors. According to the wavelet transformation analysis on the random vibration signal acquired from the acceleration sensor, the random vibration signal related to the silkworm chrysalis quality has been analyzed and reconstructed; then the characteristic values such as: mean value, variance, mean square root, waveform index, pulse factor, and so on of the quality signal have been extracted also; and then the characteristic values of the quality have been optimally selected within a fuzzy clustering method; at the end, a RBF neural network testing model with characteristic values from the silkworm chrysalis quality as an input signal was built. With these procedures, this paper has established a quality nondestructive testing system for silkworm chrysalis quality based on acceleration sensor signal measurement. The results from the application experiments demonstrated the effectiveness and applicability of this quality nondestructive system for quality testing of biological products. This quality nondestructive testing system has many advantages, including shortening the testing time, avoiding sample waste from traditional testing method, increasing the accuracy and reliability, which shows many bright social and economic benefits. This paper also provides the design and application of quality nondestructive testing systems based on vibration signal analysis with a theoretical support and experimental basis.