Using modular identification methods in gene-drug multiplex networks to infer new gene-drug associations can identify new therapeutic target genes for known drugs. In this paper, based on the gene expression data and drug response data of lung cancer in the genomics of drug sensitivity in cancer (GDSC) database, a multiple network algorithm is proposed. First, a heterogeneous network of genes of lung cancer and drugs in different cell lines is constructed, and then a network module identification method based on graph entropy is used. In this heterogeneous network, network modules are identified, and five lung cancer gene-drug association modules are identified through iterative convergence. Compared with other methods, the algorithm has better results in terms of running time, accuracy and robustness, and the identified modules have obvious biological significance. The research results in this article have guiding significance for the medication and treatment of lung cancer, and can provide references for the treatment of other diseases with the same targeted genes.