In addition to socioeconomic factors, urban morphology is increasingly being recognized for the potential role it might play in tackling climate change. Meanwhile, improving CO2 emission efficiency has become a significant aim in reducing CO2 emissions without sacrificing economic development. Hence, investigating the effects of urban landscape pattern on CO2 emission efficiency is of great importance. As thus, this study is designed to quantitatively estimate the associations between urban landscape pattern and CO2 emission efficiency, using panel data for five megacities in China in the period 1990–2013. A series of landscape metrics were first selected to quantify three dimensions of urban landscape pattern, namely urban expansion, urban shape complexity, and urban compactness. CO2 emission efficiency was then evaluated using three quantitative indicators, namely per capita CO2 emission, CO2 emission intensity, and CO2 emission performance. Finally, panel data models were performed to estimate the influences of urban landscape pattern on CO2 emission efficiency. The results demonstrate that both urban expansion and urban shape complexity inhibited the improvement of CO2 emission performance. In addition, urban compactness was found to be conducive to the reduction of CO2 emission per capita and CO2 emission intensity, as well as the improvement of CO2 emission performance. The findings obtained through this study cast a new light on the significance of city planning and the optimization of spatial structure in improving CO2 emission efficiency, providing decision makers and urban planners with a scientific basis from which to approach the task of reducing CO2 emissions whilst maintaining economic development.