This study offers a unique contribution to the literature by investigating the convergence of province-level carbon dioxide emission intensities among a panel of 30 provinces in China over the period 1990-2010. We use a novel, spatial dynamic panel data model to evaluate an empirically testable hypothesis of convergence among provinces. Our results suggest that: (1) CO2 emission intensities are converging across provinces in China; (2) the rate of convergence is higher with the dynamic panel data model than the cross-sectional regression models; and, (3) province-level CO2 emission intensities are spatially correlated and the rate of convergence, when controlling for spatial autocorrelation, is higher than with the non-spatial models. (C) 2015 Elsevier Ltd. All rights reserved.