In crop production estimation using remotely sensed data, it is necessary to estimate the crop area and its yeild per unitarea in a particular administrative division. Two approaches were adopted: (1) the target image of the study area was cut with the administrative boundary, land cover/use classification was processed and crop identification and area calculation were carried out, (2) land cover/use classification was carried out first and then the study area was cut from the image with administrative boundary, and area was estimated. We call these two methods strategy A (cut and classify) and strategy B (classify and cut), respectively. In this paper, the two strategies were applied to rice crop area estimation. The results indicate that strategy B is better than strategy A in the unsupervized classification-cluster process and the rice crop area was estimated with an accuracy over 81% for semi-late rice and 90% for early rice.