An electronic nose was used to analyze adulterated honey samples. The response signals were analyzed by principal component analysis (PCA) and partial least-squares analysis (PLS). PLS model was used to predict adulterated ratio of fructose and glucose in honey samples. The results indicated that coefficient of determination between sensors response signals and the ratio of fructose and glucose of PLS model was 0.980 3. The prediction error of PLS model was within 8% (the adulterated ratio of fructose and glucose from 20% to 70%). It was proved that electronic nose could be applied in honey discrimination when it was adulterated with high content of fructose and glucose.