A variety of genetic diseases in the retina, including retinitis pigmentosa and leber congenital amaurosis, might be excellent targets for gene delivery as treatment. A major challenge in non-viral gene delivery remains finding a safe and effective delivery system. Poly(beta-amino ester)s (PBAEs) have shown great potential as gene delivery reagents because they are easily synthesized and they transfect a wide variety of cell types with high efficacy in vitro. We synthesized a combinatorial library of PBAEs and evaluated them for transfection efficacy and toxicity in retinal pigment epithelial (ARPE-19) cells to identify lead polymer structures and transfection formulations. Our optimal polymer (B5-S5-E7 at 60 w/w polymer:DNA ratio) transfected ARPE-19 cells with 44±5% transfection efficacy, significantly higher than with optimized formulations of leading commercially available reagents Lipofectamine 2000 (26±7%) and X-tremeGENE HP DNA (22±6%); (p<0.001 for both). Ten formulations exceeded 30% transfection efficacy. This high non-viral efficacy was achieved with comparable cytotoxicity (23±6%) to controls; optimized formulations of Lipofectamine 2000 and X-tremeGENE HP DNA showed 15±3% and 32±9% toxicity respectively (p>0.05 for both). Our optimal polymer was also significantly better than a gold standard polymeric transfection reagent, branched 25 kDa polyethyleneimine (PEI), which achieved only 8±1% transfection efficacy with 25±6% cytotoxicity. Subretinal injections using lyophilized GFP-PBAE nanoparticles resulted in 1.1±1×10(3)-fold and 1.5±0.7×10(3)-fold increased GFP expression in the retinal pigment epithelium (RPE)/choroid and neural retina respectively, compared to injection of DNA alone (p = 0.003 for RPE/choroid, p<0.001 for neural retina). The successful transfection of the RPE in vivo suggests that these nanoparticles could be used to study a number of genetic diseases in the laboratory with the potential to treat debilitating eye diseases.