The purposes of this study are to examine the validity and reliability of the 2012 Environmental Performance Index (EPI) and to make suggestions for improving the EPI. The confirmatory factor analyses reveal that the constructs of environmental health, air (effects on human health), water (effects on human health), air (ecosystem effects), water resources (ecosystem effects), biodiversity & habitat, agriculture, forests, fisheries, climate change & energy have low convergent and discriminant validity. The internal consistencies of air (effects on human health), biodiversity & habitat, agriculture, forests, fisheries are also very low. Furthermore, the higher order factor analyses indicate that the fit indices of the EPI model are low. In the measurement model, child mortality, particulate matter, indoor air pollution, access to sanitation, access to drinking water, SO2 per Capita, SO2 per GDP, change in water quantity, critical habitat protection, biome protection, marine protected areas, agricultural subsidies, pesticide regulation, forest growing stock, change in forest cover, forest loss, coastal shelf fishing pressure, fish stocks overexploited, CO2 per Capita, CO2 per GDP, CO2 per KWH, renewable electricity have high measurement errors. The results suggest a bad fit to the data. Based on these findings, this study discusses some considerations to improve the EPI.