The seasonal changes of the ten-day mean sea surface temperature in the East ChinaSea for the years 1953—1972 were studied using the singular value decomposition technique.Main results obtained are as follows:In the singular vector analysis of the ten-day mean sea surface temperature in the EastChina Sea,the first three singular vectors(or seasonal march functions)account for 97.76%of the variance of the data.The geographical distribution in the weightings(or spatial pattern functions)on thesevectors shows a remarkable regional organization.An objective classification of the surface water mass in the East China Sea was basedon the mode of its seasonal temperature changes. The seasonal changes of the ten-day mean sea surface temperature in the East China Sea for the years 1953-1972 were studied using the singular value decomposition technique. Main results obtained are as follows: In the singular vector analysis of the ten-day mean sea surface temperature in the East China Sea, the first three singular vectors (or seasonal march functions) account for 97.76% of the variance of the data. geographical distribution in the weightings (or spatial pattern functions) on thesectors shows a remarkable regional organization. An objective classification of the surface water mass in the East China Sea was basedon the mode of its seasonal temperature changes.