Abstract: The gas emission features under different job types in tunneling face are studied, two dynamic indexes of fluctuation range and lower deviation which closely related with job type gas emission are put forward, and the identify threshold values of job types through analysis of samples are obtained. Accordingly, the identification model of job type is established, and a real-time dynamic recognition software modules for job types is developed. The model is applied to carry out tunneling face recognition for job types. The results are accurate and intuitive, and perfectly coincide with the real work cycle.