With the rapid development of highspeed railway, the effective detection of rail cracks for the safe operation of the railway is of great significance. Focused on the problem of rail crack recognition based on magnetic flux leakage signal, multisensor featuredecision fusion technology is adopted, which does the multisensor signal decisionfusion at the same time of the multifeature extraction and fusion of the magnetic flux leakage signal in time domain and frequency domain. And a multisensor information fusion classifier is designed based on SVM. Using the measured magnetic flux leakage signal of artificial crack, compared with the extraction of a single feature and the use of a single sensor signal to identify, the proposed method achieves a better crack recognition effect whose average recognition rate reaches 98%.