Quantum Fourier transform is the basis of quantum algorithm, and is also the key of exponential efficiency. A feature extraction algorithm based on quantum Fourier transform was proposed. In this algorithm, the operation circuit of quantum computation was built. The feature extraction steps for the execution of quantum Fourier transform were configured. The estimation function of peak was constructed for evalue the extracted features. The proposed algorithm was applied to pattern recognition of gear fault conditions. The experiment results verify the efficiency and practicability of the proposed algorithm.