The purpose of this study was to investigate the fate of transplanted cells in the central zone of myocardial infarction (MI), and to clarify the relationship between the injection-site impact and the efficacy of cell therapy. MI was created by coronary ligation in female rats. Three weeks later, 3-million labelled male bone marrow mesenchymal stem cells (BMSCs) were directly injected into the border (BZC group) or central zone (CZC group) of MI area. As a control, culture medium was injected into the same sites. Cell survival was evaluated by quantitative real-time polymerase chain reaction, and apoptosis was assayed with TUNEL and caspase-3 staining. Four weeks after transplantation, heart function and cardiac morphometry were evaluated by echocardiography and Masson's Trichrome staining, respectively. Angiogenesis and myogenesis were detected by immunofluorescence staining. After cell transplantation into the border or central zone, there was no cell migration between the different zones of MI. BMSCs in the CZC group exhibited no difference in apoptotic percentage, in the long-term survival, when compared with those in the BZC group. However, they did effectively promote angiogenesis and cellular myogenic differentiation. Although cell delivery in the central zone of MI had no effect on the recovery of heart function compared with the BZC group, the retained BMSCs could still increase the scar thickness, and subsequently exhibit a trend in the reverse remodelling of ventricular dilation. Hence, we concluded that the central zone of MI should not be ignored during cell-based therapy. Multiple site injection (border+central zone) is strongly recommended during the procedure of cell transplantation.