Moldy corn not only causes huge financial losses to our country, but also does great harm to human health and life safety. Stochastic resonance signal-to-noise ratio analysis technique is a novel method in sensor signal feature extraction field. In this paper, a gas detection method using semiconductor gas sensor and stochastic resonance signal-to-noise ratio analysis technique was established for rapid detection of moldy corn. Four kinds of corn samples were detected by the proposed method. The results showed that the signal-to-noise ratio analysis technique could characterize the moldy corn feature, and the detecting signal did not need pre-processing. The system has the advantages such as fast response, shorter desorption time, easy to use, low cost, durable, and so on. The system is promising in practical application.