Abstract: This study aims to predict candidate small-molecule drugs for the treatment of Gastric Cancer (GC) and identify their potential targets by using drug repositioning and bioinformatics methods, providing new ideas for GC therapy. In this study, 203 differentially expressed genes were first identified in GC datasets GSE54129, GSE26899 and GSE65801, and 10 candidate small-molecule compounds were screened by the Connectivity Map (CMap) analysis. Then, the topological properties of the protein-protein interaction network composed of differentially expressed genes were analyzed, and TIMP1, PDGFRB, COL1A1, etc. were identified as hub genes. Furthermore, molecular docking results showed that the newly identified small-molecule drug levonorgestrel had strong binding ability with TIMP1, and its binding energy was −9.24 kcal/mol; the binding energy of the reported GC drug Cediranib to the target PDGFRB was −6.28 kcal/mol. In addition, the expression patterns, diagnostic value and prognostic value of the potential target genes TIMP1 and PDGFRB were validated in the GC datasets.