Abstract: In recent years, due to the frequently occurred extreme weather and natural disasters, considerable amount of farmlands have been destructed, resulting in the emergence of micro-topographic features（bump features and concave features） within farmlands, thereby affecting farming. In order to solve the above problems, based on high-precision farmland digital terrain model and genetic ant colony algorithm, a method of soil allocation path planning for farmland micro-terrain features was proposed. Firstly, based on the SfM（Structure from Motion） technology to process aerial images of the test field, a high-precision farmland digital terrain model（FDTM） was obtained, and 16 bump features and 9 concave features were extracted according to the comprehensive membership degree of terrain factors, and the cut volume as well as fill volume was calculated as 0. 885 m3 and 0. 884 m3, respectively. Secondly, with the cost of earthwork reallocation as the decision target, cutting and filling areas were established as path search nodes, and the ant colony algorithm was used to obtain the initial feasible solutions.Subsequently, these solutions were preliminarily optimized by using the fitness function in the genetic algorithm. Finally, the optimal earthwork reallocation path was obtained by secondary optimization of the solutions according to the crossover and variation operation. The results showed that the global optimal solution was gained after 232 iterations, and the earthwork reallocation cost was reduced by 2. 1% compared with the traditional ant colony algorithm. The research results can provide references for precise land leveling with respect to micro-topographic features.