Abstract: To solve the issues of blind identification of primitive BCH codes encoding parameters, a novel identification algorithm with probability approximation is presented. Frist, by taking advantage of the approximation of random code words root probability character which uses Gaussian distribution and Poisson distribution, the thresholds for searching code length are structured. Second, though analyzing the checking ability of the primitive element and the impact of isomorphism on searching, the coding filed is determined by using the method of nearby fields pair which improves the performace of identification. Finally, the calculation is reduced by creating and using the conjugate roots table in the recognition of generator polynomial. Simulation results show that, the proposed algorithm achieves a significant improvement in identification probability even if in high BER situation.