Our understanding of the relationship between the neuroanatomic loci of brain damage and the incidence of post-stroke depression (PSD) is not complete. Many studies have investigated this relationship and the evidence is conflicting. With the purpose of gaining a consistent, strong, and credible conclusion on the relationship between PSD and the loci of brain damage, a meta-analysis was used in this study to systematically reanalyze the findings of related studies and to investigate the sources of heterogeneity among study results. The key words "stroke or cerebrovascular" and "depression or mood or affective" were entered into the MEDLINE, PsycINFO, and EMBASE databases to search for relevant studies. The references cited in the studies found were also used to locate additional studies. For each eligible study, the important study characteristics were recorded, and the effect sizes of the relationship between PSD and lesion location were computed. Furthermore, we conducted subgroup analyses to explore the heterogeneity among study results. A total of 3,668 patients participating in 52 studies were included in this meta-analysis. There was a weak relationship between PSD and right hemisphere lesion. The major sources of heterogeneous study results included systematic exclusion of patients with language dysfunction and use of different assessors and instruments for diagnosing depression. Future efforts should aim to enhance standards for reporting studies, improve assessment tools for assessing depression of aphasic patients, and adopt appropriate study methodologies for investigating the relationship between PSD and lesion location.