Abstract: A brain network and brain addictive connectivity computing method and device are disclosed. The method includes: obtaining real brain functional magnetic resonance (fMRI) images associated with different labels; generating first brain topologies associated with the different labels based on at least one training real brain fMRI image sample; generating a brain addiction standard feature map based on an additive real brain fMRI image; determining an initial weight value of each of the first brain topologies according to the brain addiction standard feature map; training an addictive brain network analysis model based on the first brain topologies associated with different labels and the initial weight value of each first brain topology; inputting at least one testing real brain fMRI image sample into addictive brain network analysis model to generate first weighted brain topologies.