In this paper, two cancer recognition models, global component model (GCM) and cancer component model (CCM), are constructed. Due to the fact that GCM and CCM complement each other, a weighted voting strategy is applied, and an ensemble algorithm based on GCM and CCM for cancer recognition (EAGC) is proposed. Independent test experiments and cross validation experiments are conducted on Leukemia, Breast, Prostate, DLBCL, Colon, and Ovarian cancer dataset, respectively, and EAGC performed well on all datasets. The experimental results show that recognition, solution, and the generalization are strengthened by the combination of GCM and CCM.