Abstract: In order to realize blind recognition of BCH codes in the case of unknown channel error probability and imperfect synchronization, an improved blind recognition algorithm was proposed. Firstly, estimated channel error probability by combining modulation scheme and SNR estimation algorithm. Then, deduced an optimal decision threshold for deciding whether a minimal polynomial under the testing field index was a factor of the generator polynomial or not according to the estimated channel error probability, and completed the recognition of the code length. At last, recognized the real synchronization deviation by comparing the number that the row polynomials were divisible by the minimal polynomial under each tested synchronization deviation, and completed the recognition of the generator polynomial by finding all the minimal polynomials that satisfied the decision threshold. Simulation results demonstrate that the proposed algorithm can recognize the BCH codes efficiently in the case of unknown channel error probability and imperfect synchronization, and the recognition performance outperforms the previous algorithms.