Detecting the atmospheric CO2 reliably is becoming more and more important due to environment problems and climate changing. Raman light detection and range (lidar) system monitoring the atmosphere is a novel technique. It is a key factor to acquire the back signals of Raman lidar efficiently. The paper not only analyzes the principle of Raman lidar and the theoretical basis of signal processing, but also designs two schemes for collecting Raman signals, which include single channel and double channels, and demonstrates the methods of signal processing separately. The paper resolves the noise interference in collecting signals process, and gets the satisfied conclusion by contrasting the two results of the two schemes. Moreover, the signal-to-noise ratio (SNR) is calculated for the lidar system. The calculated results indicate that the SNR (for 1 km) is up to 10 for single channel, however, up to about 20 for double channels.