In response to the problem that existing methanol gas detection methods cannot meet the rapidly growing demand for rapid detection in the fuel cell market, a detection system design and quantitative analysis were conducted focusing on methanol gas. Firstly, the light source was selected based on the infrared absorption characteristics of methanol gas, and its driving circuit was designed. The gas modulates the light source. Secondly, an orthogonal lock-in amplification system was designed using wavelet decomposition and lock-in amplification techniques. The lock-in amplification output signal was further processed using wavelet decomposition, achieving a detection limit of 9 ppm and a response time of approximately 2 s, enabling the detection of weak signals. Lastly, an analysis model was established based on linear regression with a signal determination coefficient of 0.979. Experimental results demonstrate that the designed methanol detection system has accurate predictive capabilities and high sensitivity, enabling rapid and quantitative detection of methanol.