Fuente: WIPO "processing tomato" The invention relates to the field of tomato picking, in particular to a clustered tomato picking robot picking strategy planning method based on self-adaption 26 neighborhoods, and aims to overcome the defects in the background technology so as to effectively improve the clustered tomato picking success rate. According to the technical scheme, the method comprises the following steps: S1, obtaining tomato and stem information, and obtaining a tomato three-dimensional pose and a stem three-dimensional position through a multi-task convolutional neural network and a point cloud processing algorithm; s2, inter-cluster picking planning is carried out, inter-cluster picking sequence planning is carried out through a DBSCAN clustering algorithm, and an optimal picking cluster is obtained; s3, in-cluster picking planning is carried out, self-adaptive 26 neighborhoods of the target tomatoes are established according to manual picking experience and mechanical picking experience, and an in-cluster picking sequence planning algorithm is provided; s4, picking attitude planning: putting forward a picking attitude planning algorithm according to distribution of tomatoes and stalks in self-adaptive 26 neighborhoods; s5, the clustered tomatoes are picked one by one.