In contrast to the conventional conditional mean approaches, this paper provides the first application of the quantile regression method, with a focus on unbalance panel data application, to re-examine the relationship between CO2emissionsand income across different quantiles of the conditional CO2 emissions function. The empirical findings support theinverted-U hypothesis between two variables at most conditional quantiles, and these results proverobust to different controlling variables as well as parametric model specifications. One of the main results that suggest positive income shocks have stronger impact on CO2 emissions vary across the quantiles, indicators of environmental degradation first rise, and the fall with increasing economic development. Moreover, each quantiles in both... case lead to larger marginal effects of energy use on CO2 emissions. In line with the literature, the estimates of the energy variable are each positive, statistically significant, and large in magnitude. The findings from our more flexible approaches indicate that there are overwhelming evidences in support of the CO2 emissions is mainly determined by energy use.