Abstract: [Objective]As one of the major oil crops, oilseed rape has an important position in the national economy. The fragmentation of rape planting in our country significantly increased the difficulty of monitoring its growing area, as the regular ground survey takes much time and energy with low representation.Remote sensing, with the features of covering large areas and revisiting in a high frequency, has become the most ideal method for monitoring rape growing area. The common remote sensing monitoring methods including visual interpretation and supervised classification are comparatively more influenced by the subjectivity of the monitors, due to the high degree of human-computer interaction.Aiming at the above problems; this article proposes a new monitoring method on the basis of unsupervised classification after layering. [Method]The GF images of the year 2013 in Jiang County have been adopted for application and precision evaluation by equidistance layering and natural layering. [Result]The result shows that the unsupervised classification after layering is higher in the overall precision than the direct unsupervised classification. The overall precisions on the basis of equidistance layering and natural layering increased from 79.16% to 84.44% and 85.17% respectively. The user precision in the direct unsupervised classification is only 72.97%. After equidistance layering and natural layering, the precision rate have been raised up to 81.05% and 86.12% respectively, leading to a great reduction of misjudging the non-rape area as the rape area. The overall precisions of equidistance layering and natural layering are 84.44% and 85.17%, with the Kappa coefficient to be 0.69 and 0.70. There is no significant difference in the precision between two layering methods. But the natural layering has a comparatively higher degree of user precision, mapping precision and overall precision with higher reliability. [Conclusion] The new methods proposed in this article has little manual intervention and high precision, which has a large potential to be applied in area monitoring on the basis of mass images.