The honey quality could be predicted based on the electrical conductivity that dissolves in honey. Currently, dissolved electrical conductivity could be measured only using a traditional instrument, namely a conductivity meter. A measurement system based on visual near-infrared imaging was developed to predict the electrical conductivity value of the honey. The reference of the electrical conductivity was standardized using a conductivity meter, namely HI 98311 that produced by Hanna Instrument, Rhode Island, USA. The measurement benchwork was build using 40×40 aluminum profiles to handle a hyperspectral camera, a slider system to move the object that put-on the Teflon sheet. The honey sample was captured in transmittance mode using a hyperspectral camera with a wavelength range of 400-1000 nm. The sample was captured in 448 bands of wavelength with a 1.3 nm step. The illumination box was developed to transmit light from a halogen lamp to a honey sample in transmittance mode. The data processing consists of image correction, feature extraction, wavelength selection, and prediction model. The wavelength selection and prediction model are built using a combination of Partial Least Square and Artificial Neural Network (PLS-ANN). Hypercube images were prepared from 28 honey typed to evaluate the performance of the proposed system. The performance of PLS-ANN model represented by a linear correlation coefficient is 0.98, whereas the root means square error (RMSE) is 12.6%. Finally, the proposed method based on visible near-infrared imaging could be used to determine the electrical conductivity containing in honey non-destructively.