With the aim to solve the problem of low detection speed and precision of beef tenderness, a fast nondestructive detection method for beef tenderness based on airpuff and structural light 3D imaging technology was proposed. The structural light 3D scanning technology was used to obtain the threedimensional point cloud information on the surface of the beef and the point cloud processing technology was combined to extract the depth, area, surface area and volume parameters of the stressed depression area on the beef. In point cloud processing, denoising, point cloud segmentation, greedy projection triangulation, Delaunay triangulation, surface fitting and other algorithms were used to extract the characteristic parameters of beef samples. The modeling method was used to establish the prediction model of beef shear force which about the least squares support vector machine regression (LS-SVR), back propagation (BP) and general regression neural network (GRNN). The results showed that the GRNN model performed the best with the correlation coefficients of prediction set of 0.975, and root mean squared error of 5.307N. The GRNN neural network based on K-fold cross validation was used to predict the tenderness grade. It was worth noting that the results showed that the method had a better grading effect on the tender beef of 100% and a slightly worse grading effect on the tougher beef of 91.3%. The results demonstrated that the proposed airpuff combined with structured light method was effective in beef tenderness detection nondestructively. The research result provided a method for poultry meat tenderness detection and a basis for online poultry tenderness detection which had broad application prospect not only in meat tenderness but also in fruit hardness and ripeness detection.