The invention provides a non-common-view multi-sensor fusion method based on label Dobernoulli distribution, and belongs to the field of multi-sensor fusion. According to the non-common-view multi-sensor fusion method provided by the invention, the fusion posteriori distribution of nodes is compensated by uniform distribution, and the fusion of the posterior information of multiple non-common-viewnodes is realized; a network consensus strategy is adopted, and the fusion of global posteriori distribution of the nodes of a sensor network is realized by using the posteriori information between the nodes and the transmission of identification information; and by means of joint grouping of target tracks between the sensors, the operating rate of an inter-group LF-GCI fusion algorithm is increased by 1 to 1.5 orders of magnitude, the non-common-view multi-sensor fusion method provided by the invention is applicable to the fusion of the posterior information of any common-view and non-common-view layout forms of the sensor network, the algorithm execution efficiency is high, and the non-common-view multi-sensor fusion method is robust to the "label inconsistency" sensitive problem between the sensor nodes.