The invention discloses a method and device for classifying cross domain viewpoint data, and belongs to the technical field of the Internet. The method comprises: obtaining a shared topic matrix according to a shared topic of a source domain and a target domain, and respectively obtaining a specific domain topic matrix of the source domain and a specific domain topic matrix of the target domain according to a specific domain topic of the source domain and a specific topic of the target domain; determining an objective function of the source domain, and determining an objective function of the target domain; determining a total objective function according to the objective function of the source domain and the objective function of the target domain; determining convergency values of parameters in the total objective function, and obtaining classification functions according to the convergency values of the parameters in the total objective function; and classifying the viewpoint data of the target domain according to the classification functions. According to the method and the device disclosed by the invention, the cross domain viewpoint data are classified according to the classification functions obtained by the shared topic matrix, and since the shared topic matrix can reduce the difference of different domains, the classification precision of the cross domain viewpoint data is improved.