The potential of using hyperspectral imaging in the wavelength region of visible and short-wave near infrared range (400-1000 nm) was investigated for the rapid and non-invasive determination of mechanical properties in different parts of prawns (Metapenaeus ensis). Hyperspectral cubes were acquired at the second and fourth segments of prawn and their mean spectral data were extracted. The quantitative relationship between the spectral data and their corresponding reference mechanical properties (measured by Instron universal testing machine) were established by partial least square regression (PLSR) and least-squares support vector machines (LS-SVM) algorithms, respectively. Successive projections algorithm (SPA) was carried out to select the most important and effective wavelengths, which was an optimization process for improving the performance of established models. Then the optimized models were built using PLSR and LS-SVM based on the selected wavelengths and their performances were compared to find the best model for predicting mechanical properties of prawn. As a result, the SPA-LS-SVM model was considered as the best predictive model to determine the hardness, gumminess and chewiness values of prawn with correlation coefficient of prediction (R-p) of 0.8489, 0.8096, and 0.8596 and root mean square error of prediction (RMSEP) of 0.1465, 0.1445, and 0.1258, respectively. Visualization of distribution maps for the mechanical properties within the region of interests (ROI) of prawns was also presented. The overall results revealed that hyperspectral imaging technique had a great ability for predicting the mechanical properties of prawn rapidly and non-invasively. (c) 2014 Elsevier Ltd. All rights reserved.