Traditional manipulation excitation of fight data recognition methods rely on manual analysis and judgment, which require significant effort and have low efficiency. The manipulation excitation recognition technology based on short-time frequency analysis proposed in this paper involves converting flight test data into the time-frequency domain for analysis using the Short-Time Fourier Transform (STFT) algorithm, then extracting feature information of manipulation excitations of the data using Sobel operators. This enables automatic recognition of manipulation excitations, significantly improving the efficiency and accuracy of flight test data analysis. The application of this technology can better understand and grasp the performance and characteristics of aircraft, providing important data support for aircraft design and improvement.