As to the complexity and low analysis efficiency of defect identification inverse problem solution by manual programming model, with the conjugate gradient algorithm, a method for the secondary development of the universal finite element numerical computation software Ansys was put forward in this paper, which is successfully applied to the diagnosis of the irregular inner-wall surface defect by infrared inspection. And a new concept of relative sensitivity was introduced here in order to assess the influence of inspection state to the quantitative identification of defect boundary shape. Through analyzing the relative sensitivity, it was found that the inspectability of the irregular inner-wall surface defect was different in the different inspection states, and the inspection carried out at the time with maximum inspection surface temperature difference was more desirable than that in the steady heat transfer state. Both of the feasibility of the infrared diagnosis with Ansys secondary development and the validity of the relative sensitivity assessment were verified by the numerical experiment.