Abstract: Abstract: To the corresponding research on the problem of TDDM-BOC（Time Division Data Modulation-Binary Offset Carrier）modulation signal under low signal-to-noise ratio, including blind estimation of the combination code sequence, the article proposes a multi principal component neural network (Sanger NN) . Firstly, the segmented TDDM-BOC signal is used as the input signal and we used Sanger NN to extract adaptively weight vector of multi-feature component. Then the weight vectors of neural network through the continuous input signal are trained repeatedly until convergence. Finally, the signal of the combination code sequence can be rebuilt by the symbolic function of each weight vector, thus realizing the blind estimation of the TDDM-BOC signal. Furthermore, an optimal variable step convergence model is used in this paper, which improves greatly the convergence speed of the neural network. Theoretical analysis and simulation results show that the Sanger NN method can achieve the combined code blind estimation of TDDM-BOC signals under low signal-to-noise ratio about . Compared with singular value decomposition (SVD), it is significantly reduced for the complexity of Sanger neural networks used in the TDDM-BOC signal of combination code blind estimation.