The changes in the aroma and spoilage odor emitted from eleven ‘Hongro’ apples during ten weeks’ storage were investigated using six types of metal oxide semiconductor gas sensor arrays. The gas sensors used in the evaluation were sensitive to apple-emitted aroma or spoilage odor, and a high reproducibility of 5% relative standard deviation or less was confirmed. Significantly, the change in apple-emitted aroma or spoilage odor was easily distinguished by the optimal gas sensor and a significant correlation (r=0.992) between decay rate and sensitivity change was observed. The results of a principal component analysis of the signal patterns obtained by data standardization using the optimal gas sensor showed a clear classification between decayed sampler groups and undecayed sampler groups.