Fronts in the Bohai sea are detected based on BJ-1 small satellite data using a gravitation model.We select the red band as the experimental band due to the characters of small satellite data. In addition, the fronts south of the Bohai Bay and east of the Liaodong Bay are chosen as experimental areas because they are both eligible for front detection. In the experiments, a linear stretch and band computation are first applied to data in order to protrude front information. The stretch intervals are set according to histograms. Then, utilizing a gravitation model, the fronts can be extracted from preprocessed data,and the results show that fronts can be successfully extracted from BJ-1 small satellite data usingthe gravitation model. The detected fronts are continuous and obvious. Compared with traditional front detection algorithms, the proposed algorithm can not only detect continuous and conspicuous fronts, but can also suppress noises.