Nowadays, the vigorous development of artificial intelligence technology and connotation not only injects new vitality into Chinese movies, but also promotes the exploration and development of new paths of cultural dissemination of Chinese movies. In this paper, based on graph convolutional neural network and heterogeneous network, M-DGCN, a model of cultural communication of Chinese movie industry, is established. it utilizes Motif structure to obtain the higher-order spatial information of the network data of cultural communication of Chinese movie industry and combines with the two-channel graph stochastic convolutional network to highlight the more valuable information in the process of cultural communication of Chinese movies. The qualitative comparative analysis method is then introduced to explore the factors affecting the development of cultural communication power of Chinese film industry, using IMDB database as the source of research data. The study shows that the MAE error of the M-DGCN model tends to decrease with the increase of the number of nearest-neighbor layers considered, and the consistency indices of the conditional groupings in the QCA results exceed the threshold value of 0.85, which indicates the stronger explanatory power of the path combinations for the samples. Therefore, the optimization of the cultural communication mode of Chinese film industry in the intelligent era needs to be based on the dimensions of content, marketing and experience in order to better promote Chinese films to the world.
