This study aims to investigate the impact of foreign direct investment (FDI), together with gross domestic product (GDP), the square of GDP, and energy consumption, on carbon dioxide (CO2) emissions in Turkey over the period 1974-2010. We employ both the bounds test approach which has superior properties especially in small samples and the Hatemi-J test which takes structural breaks into consideration in the co-integration analysis. Due to the co-integration relationship between CO2 emissions and other variables, the autoregressive distributed lag (ARDL) model is used in order to investigate short and long run elasticity between the variables. The long-run coefficients of the ARDL model indicate that the effect of FDI on CO2 emissions is positive but relatively small, while the effects of the GDP and energy consumption on CO2 emissions are quite considerable. Moreover, the short-run coefficients obtained by the error correction model (ECM) are found to be similar to those of the long-run model. The findings support the validity of the environmental Kuznets curve (EKC) hypothesis in both time-horizons. The vector ECM based Granger causality test is also applied to investigate the causal link. The causality test results indicate the existence of a causality running from all explanatory variables to CO2 emissions in the long run. Overall, the findings suggest that Turkey should promote energy efficiency with sustainable growth, and encourage more FDI inflows particularly in technology-intensive and environment-friendly industries to improve environmental quality. (C) 2015 Elsevier Ltd. All rights reserved.