CO2, as the main component of greenhouse gases, has a major impact on the global climate. Currently, passive observation equipment can only get the CO2 concentration over a single point. However, differential absorption lidar has great advantages in the detection of CO2 and can get higher spatial and temporal resolution. At present, when using the differential absorption lidar to detect the concentration of CO2, the pressure broadening cannot be ignored. Focusing on the analysis of pressure broadening, a new correction algorithm is proposed which reduces the error in the inversion process when compared with the classical algorithm. After the inversion of the analog signal and measured signal, ultimately, the ideal results of the vertical profile of CO2 concentration distribution are got and the feasibility of the new correction algorithm is verified.