Spoilage spawned by Zygosaccharomyces rouxii can cause sensory defect in apple juice, which could hardly be perceived at the early stage and therefore would lead to the serious economic loss. Thus, it is essential to detect the contamination earliest to avoid costly waste of products or recalls.The present work tried to explore the feasibility of detection of the spoilage in apple juice caused by Zygosaccharomyces rouxii (Z.rouxii) at early stage through determination of flavor of apple juice by using electronic nose (EN). The ability of EN to discriminate apple juices with different contamination time was analyzed by using linear discrimination analysis (LDA) method. The contribution of each sensor to the diagnosis of the Z. rouxii\|contaminated apple juice by EN was studied using loadings analysis. Test panel evaluation was performed as a reference to validate the performance of EN technique.The correlation between electronic nose signals and yeast concentration was analyzed by partial least squares analysis (PLS) method. The results showed that EN was able to accurately identify the Z.rouxii\|contaminated apple juice after incubation for 12 h,corresponding to the fungal number of less than 200 CFU/mL, the level at which the test panel could not yet identify the spoilage.Loadings analysis indicated that sensors S1, S2, S6, S7 and S8 had the most important effect on identification of apple juice contaminated at different levels.A good correlation between sensors response signals and yeast concentration was revealed by PLS algorithm,with correlation coefficient R2 of 0.97 and 0.93 in calibration model and prediction model, respectively. The standard error of calibration and standard error of prediction were 0.28 and 0.26, respectively.These results indicate that it is feasible to detect the spoilage in apple juice caused by Z. rouxii at early stage using electronic nose, and this sensory technique is expected to be applied widely in the field of detection of foodborne microorganisms.