This study reappraises the relationships between financial development and carbon dioxide emissions by using 25 OECD countries during 1971–2007 as observations. It introduces the panel transition regression (PSTR) model. We found that strong evidence of the relationship between financial development and carbon dioxide emissions is non-linear and the trade-off correlation between these ratios and the carbon dioxide emissions. The carbon dioxide emissions will be different under the financial development threshold value and the control variables of energy consumption, GDP and GDP2. What is more, the different financial development attributes produce completely different carbon dioxide emissions. In sum, the threshold effect of financial development will be an important index to control carbon dioxide emissions.