Air-insulated switchgear cabinet plays a critical role in entire power transmission and distribution system. Its stability directly affects the operational reliability of the power system. And the on-line gas detection method, which evaluates the insulation status of insulation equipment by detecting the decomposition components of filled air in cabinet, becomes an innovative way to ensure the running stability of air-insulated switchgear cabinet. In order to study the characteristic gas types and production regularity of decomposition components under partial discharge, three insulation defects: needle-plate, air-gap and impurity defect are proposed to simulate the insulation defects under partial discharge in air-insulated switchgear cabinet. Firstly, the generation pathways and mechanism of composition components are discussed. Then CO and NO2 are selected as the characteristic decomposition components to characterize the partial discharge due to their high concentration and chemical stability. Based on the different change regularity of CO and NO2 concentration under different insulation defect, it provides an effective way to evaluate and predict the insulation defect type and severity in the field.