Abstract: To handle the high delay problem of blind decoding process for blind recognition of Channel Codes, a fast blind recognition algorithm of channel codes based on maximum likelihood decoding is proposed. This algorithm combines channel decoding with channel recognizing by using Euclidean distance of maximum likelihood decoding as the recognition feature, It can recognize and terminate invalid decoding in advance during channel decoding to accelerate the decode process. The optimum detection threshold satisfying system reliability requirements is deduced by Neymar-Pearson (NP) criterion. The algorithm is uncomplex and easy to implement. Simulation results prove the validity of deduction and show that the proposed algorithm has good performance in blind channel coding recognition.