The work aims to apply finite element method to detect the surface cracks of TC4 curvature structure based on the principle of ultrasonic infrared thermography and investigate effects of excitation conditions on the detection effect to obtain the optimal excitation parameters. Firstly, several schemes consisting of different excitation amplitudes, different excitation frequencies, different excitation time and different excitation positionswere designed for simulation calculation. Secondly, the surface crack of aero-engine blade was detectedbased on the optimal excitation scheme to verify accuracy of the simulation. During the ultrasonic excitation, the temperature difference between the defect region and the non-defect region increased gradually and the maximum surface temperature difference was reached at the end of the excitation. The maximum surface temperature difference increased gradually and the increasing rate gradually decreased with the increase of excitation amplitude. However, the maximum surface temperature difference increased gradually and the increasing rate gradually increased with the increase of excitation frequency. The maximum surface temperature difference increased gradually and tended to be constant after 80 ms with the increase of excitation time. The detection effect was best when the ultrasonic excitation was applied to the center of the component in the direction of crack propagation. The detection effect was poorer when the ultrasonic excitation was applied just below the crack in the direction perpendicular to the crack propagation direction, but the effect was best when the excitation source was moved to the side for 10~20 mm. From the detection for surface crack of aero-engine blade, the crack information could be clearly seen and the good detection effects indicated the simulation results were reliable. The research results effectively reveal the impact of excitation amplitude, excitation frequency, excitation time and excitation position on the detection effect and lay a theoretical foundation for the detection optimization of surface crack detection of curvature structure by ultrasonic infrared thermography.