Abstract: Abstract: Tea is the most consumed natural plant drink in the world, and it plays an important role in human's daily life. The spatial distribution information of tea plantation is helpful for the government management and decision-making. Songyang county is located in southwest part of Zhejiang province, China, and the topography is characterized by basin in the central and surrounded by hills and mountains. The humid and cloudy climate is very suitable for tea planting, which accounts for the large proportion of tea plantation area, 68.5% of county’s whole cultivated land. In this paper, ZhangXi Town, ZhaiTan Town, YeCun Town and ZhuYuan Town of SongYang County in Zhejiang Province were chosen as study area, and ZY-3 satellite images acquired on December 25, 2012 and June 9, 2013 were used to study the method of tea plantations extraction. Eight categories including roads, water, buildings, shadows, bare soil, forest, other crops and tea plantation were identified after conducting visual interpretation and field surveys. The decision tree method was adopted to extract the tea plantations. Due to the fact that tea plants in plain areas and mountains areas show different characteristics in their planting patterns, planting area and growth status, ,the decision trees were built separately for these two different areas.The threshold values in the decision tree were determined by gradually changing their values in a certain range. Spectral curve analysis shows the range of the difference between band4 (0.77-0.89 μm) and band3 (0.63-0.69 μm) on December 25, 2012 is 20-30. The normalized difference vegetation index (NDVI) is almost unchanged or decreased from summer to winter for forest lands as they are covered mainly by evergreen broad-leaved forest, deciduous broad-leaved forest, bamboo forest and mixed forest. As for tea plant, due to its seasonal harvest and pruning in summer, NDVI in summer is a little lower than that in winter and the threshold value of NDVI difference between summer (June 9, 2013) and winter (December 25, 2012) was 0~0.1. As tea plants are terraced planted along the contour in mountain area, texture features characterized with nearly parallel line trend for tea plantations are presented in the image. The panchromatic data on December 25, 2012 was used to derive texture features. Anisotropic strength with a range of 0 to 1 was obtained after conducting the anisotropic strength algorithm. The classification results with different threshold values were compared with region-of-interest data and threshold values with the highest overall accuracy and Kappa coefficient were selected as final threshold. For plain areas, the difference between band4 and band3 was used to roughly exclude roads, water, buildings, bare soil, other crops and part of the forest from tea plantations with the value above 26. Then the threshold value of 0 for NDVI difference between summer and winter was adopted to exclude the remaining forest. Spectral feature and textural feature were both used to extract tea plantations in mountainous areas. The threshold value of 20 for band4 and band3 difference and 0 for NDVI difference between summer and winter were firstly adopted to exclude water, buildings, crops, roads, bare soil and part of forest. And the threshold of 0.35 for anisotropic strength was then adopted to exclude the remaining forest. The classification maps were validated with ground verification data and compared with results derived from neural network (NN) classification. The results show that decision tree method combining with spectral and textural information can significantly improve the classification accuracy. The overall accuracy and Kappa coefficient in the plain area were 95.00% and 0.85, respectively, increased by 5.46% and 0.19 when compared with NN classification. The overall accuracy and Kappa coefficient in the mountain area were 92.97% and 0.69, respectively, increased by 7.57% and 0.61 when compared with NN classification. The presented study could provide a reference for government forecasting crop production and preventing disaster for tea plantation.