Transplanting hematopoietic and peripheral blood-derived stem/progenitor cells can have beneficial effects in slowing the effects of heart failure. We investigated whether human bone marrow CD133 + -derived cells (BM-CD133 + cells) might be used for cell therapy of heart injury in combination with tissue engineering. We examined these cells for: 1) their in vitro capacity to be converted into cardiomyocytes (CMs), and 2) their potential for in vivo differentiation when delivered to a tissue-engineered type I collagen patch placed on injured hearts (group II). To ensure a microvascular network ready for use by the transplanted cells, cardiac injury and patching were scheduled 2 weeks before cell injection. The cardiovascular potential of the BM-CD133 + cells was compared with that of a direct injection (group I) of the same cells in heart tissue damaged according to the same schedule as for group II. While a small fraction (2 ± 0.5%) of BM-CD133 + cells cocultured with rat CMs switched in vitro to a CM-like cell phenotype, in vivo—and in both groups of nude rats transplanted with BM-CD133 + —there was no evidence of any CM differentiation (as detected by cardiac troponin I expression), but there were signs instead of new capillaries and small arterioles. While capillaries prevailed over arterioles in group II, the opposite occurred in group I. The transplanted cells further contributed to the formation of new microvessels induced by the patch (group II) but the number of vessels did not appear superior to the one developed after directly injecting the BM-CD133 + cells into the injured heart. Although chimeric human–rat microvessels were consistently found in the hearts of both groups I and II, they represented a minority (1.5–2.3%) compared with those of rat origin. Smooth muscle myosin isoform expression suggested that the arterioles achieved complete differentiation irrespective of the presence or absence of the collagen patch. These findings suggest that: 1) BM-CD133 + cells display a limited propensity for in vitro conversion to CMs; 2) the preliminarily vascularized bioscaffold did not confer a selective homing and differentiation advantage for the phenotypic conversion of BM-CD133 + cells into CMs; and 3) combined patching and cell transplantation is suitable for angiogenesis and arteriogenesis, but it does not produce better results, in terms of endothelial and smooth muscle cell differentiation, than the “traditional” method of cell injection into the myocardium.