Total volatile basic nitrogen (TVB-N) content is an important index in evaluating the chicken′s freshness. We attempt to use synergy interval partial least square coupled with genetic algorithm to select the best wavelengths. Texture features of gray images of the corresponding wavelengths are extracted. Principal component analysis (PCA) is implemented on these feature variables from image information. We take the best principal component factor numbers as the input layer. And a prediction model of the TVB-N content is developed by the back-propagation artificial neural network (BP-ANN). The results of the model are achieved as root-mean-square error (RMSE) of 6.61 and 9.84, and correlation coefficient r of 0.9054 and 0.8030 in training and prediction sets, respectively. This work demonstrates that hyperspectral imaging technique is a valid means for quick and nondestructive detection of TVB-N content in chicken.