This paper studies the optimal placement problem of biaxial acceleration sensor used for structural vibration measurement of transmission tower. On the basis of TMAC method, a novel method called Redundancy Reduced Biaxial Modal Assurance Criterion (RRBMAC) method is proposed to realize the optimal placement of biaxial acceleration sensor. The method uses improved twodimension column pivot QR factorization to gain initial sensor positions. The objective function is established based on the maximum value of nondiagonal element in BMAC matrix and measurement point redundancy, and optimization is carried out. EFI2 method and RRBMAC method were adopted to conduct the optimal sensor placement of a steel tube tower. Modal Assurance Criterion and Maximum Singular Value Decomposition Ratio along two horizontal directions were adopted to evaluate the achieved sensor placement scheme. The results show that the optimization effects of the two methods both firstly increase with the number of the sensors, and then tend to stable. RRBMAC method has better placement effects than EFI2 method and RRBMAC method can obtain pretty good optimization capability even when the number of sensors is few. In addition, the two methods were used to carry out the optimal sensor placement of a lattice transmission tower, and the comparison and analysis results show that the RRBMAC method can obtain better placement scheme. The optimal method should be chosen according to the number of sensors together with the focal point of the test. For different number of sensors, the optimal sensor placement method and some sensor placement schemes are presented. Windinduced vibration response experiment on the tower was carried out using the optimized sensor placement scheme, and NExTERA method is adopted to identify the natural frequencies and damping coefficients. Compared with the modal analysis results, the results show that the NExTERA method can extract accurate frequency results.