Background Esophageal carcinoma (EC) and gastric cardiac adenocarcinoma (GCA) have high incidence rates in the Chaoshan region of South China. Multifocal esophageal and cardiac cancer (MECC) is commonly observed in this region in clinical practice. However, the genomic characteristics of MECC remains unclear. Materials and methods In this study, a total of 2123 clinical samples of EC and GCA were analyzed to determine the frequency of multifocal tumors, as well as their occurrence sites and pathological types. Cox proportional hazards regression was used to model the relationship between age, sex, and tumor state concerning survival in our analysis of the cohort of 541 patients with available follow-up data. We performed whole-genome sequencing on 20 tumor foci and 10 normal samples from 10 MECC patients to infer clonal structure on 6 MECC patients to explore genome characteristics. Result The MECC rate of EC and GCA was 5.65% (121 of 2123). Age and sex were potential factors that may influence the risk of MECC (p < 0.001). Furthermore, MECC patients showed worse survival compared with single tumor patients. We found that 12 foci from 6 patients were multicentric origin model (MC), which exhibited significant heterogeneity of variations in paired foci and had an increased number of germline mutations in immune genes compared to metastatic model. In MC cases, different lesions in the same patient were driven by distinct mutation and copy number variation (CNV) events. Although TP53 and other driver mutation genes have a high frequency in the samples, their mutation sites show significant heterogeneity in paired tumor specimens. On the other hand, CNV genes exhibited higher concordance in paired samples, especially in the amplification of oncogenes and the deletion of tumor suppressor genes. Conclusions The extent of inter-tumor heterogeneity suggests both monoclonal and polyclonal origins of MECC, which could provide insight into the genome diversity of MECC and guide clinical implementation.