In this paper, beef mixed with a variety of spices were fried to different degrees. The physical and chemical characteristics, such as moisture content, color and tenderness, and the spectra in 370~1 023 nm were measured. Then the quality detection models and visualization method were constructed. The results showed that the moisture content decreased, and the shear force increased during the frying process, and the change was significant (P＜0.05). There was no significant difference in color between the samples and visible spectrum. While NIR spectra (800～1 023 nm) increased with cooking, and the changes were significant. Based on 100 beef data, by using Principal Component Analysis to reduce the spectral data and Support Vector Machine algorithm to establish regression model, optimizing the parameters with Particle Swarm Optimization, the values of prediction R² of beef moisture and shear force were 0.908, 0.763, and RMSEP were 1.096 and 2.097, respectively. This experiment verified the detection ability of hyperspectral imaging technology for complex food mixed with various spices. And it provides an early exploration for the intelligent production and quality monitoring of prepared meat.