The invention relates to a branch tangent line optimization setting method based on a polytrichum foraging algorithm, which comprises the following steps: firstly, connecting adjacent dipole pairs caused by noise according to the distribution characteristics of residual points in an interference fringe pattern, and grounding monopoles close to boundaries to achieve the purpose of the same number of positive and negative residual points in the whole interference pattern; Secondly, establishing a complete branch tangent line network between residual points by establishing a polytrichum gloeosporioides foraging algorithm model; Secondly, introducing a random disturbance model, and under the condition that the charge quantity of the system is kept balanced, disconnecting a redundant branch tangent line to construct a minimum spanning tree branch; And finally, bypassing the branch tangent line by using a flood overflow flooding method to carry out phase unwrapping. The method is reasonablein concept and complete in model, the total length of branch cutting can be remarkably reduced, the'islanding 'phenomenon in the unwrapping process can be effectively overcome, optimal setting of branch cutting is achieved, and the method has the advantages of being extremely high in robustness, excellent in solving quality and the like.