In this study, satellite sea surface temperature (SST) dataset is used to study the wakes formed by the Kuroshio flows past Green Island, a small island located off southeast of Taiwan. Satellite thermal Infrared Imagery usually has missing data due to cloud coverage. This situation is difficult for further analyzing. To fill up the missing data and to obtain long-term information, Data Interpolating Empirical Orthogonal Functions is used in the study. These SST images without missing data are then analyzed by the Empirical Orthogonal Function (EOF) method. The first EOF mode explains 99.95% of the total variance and clearly shows a cooling area behind Green Island. Using the gradient EOF method, each mode indicates different wake features, including direction, range, and extended length of the wake. During the data period, the percentage of the wake occurrence was 59%. The average cooling SST was about 0.91±0.15°C. For the extended range of the island wake, there is 63% shorter than 20 km, and only 8% longer than 30 km. In addition, the wake area has 1-3 times higher than surrounding area in chlorophyll-a concentration due to the island mass effect.