This work aims to develop a prediction system based on hyperspectral imaging to measure electrical conductivity value of honey. The reference of electrical conductivity was measured using a conductivity meter (HI 98311, Hanna Instrument, Rhode Island, USA) as an aqueous solution (20 g dry matter in 100 ml distilled water). The data processing consist of image correction, feature extraction, and feature selection, while as the prediction model was built using Partial Least Square Regression (PLSR), Artificial Neural Network (ANN), and Partial Least Square Artificial Neural Network (PLS-ANN). Hyperspectral images from 28 honey samples were captured using visible near-infrared (VIS-NIR) hyperspectral camera (400-1000 nm) in transmittance mode. The performance of the prediction system provided by PLSR is 0.88 (correlation coefficient) and 26.20% (RMSE). The ANN has correlation coefficient and RMSE of 0.93 and 23.12%, respectively. The best performance was the PLS-ANN model which has correlation coefficient of 0.98 and RMSE of 12.65%.