A novel algorithm named united method is presented for optimal sensor placement. The algorithm takes full advantage o f merits of both effective independence(EI) and modal assurance criterion(M AC) methods,and resolves the problem existing in the MAC method t o locate the optimal sensor placement. Firstly, to acquire the initial measuring poi nt,the technique based on a column principal component QR decomposition is adopt ed . Secondly, the candidate points to be added are determined by using the effecti ve independent method, which avoids the blindness and random of adding points. Thirdly, some candidate points are added to the measuring poin t cluster and the chosen points should meet the demand of the minim um nodiagonal element. The last step will not break until the required measuring points are obtained. T he algorithm is coded by using Matlab 70. A cablestayed bridge is chosen to test the mixed method and several o ptimal sensor placement projects with different measuring points are offered. The results show that the mixed method can not only ensure the minimum nonedia gonal component value of the MAC matrix but also ensure the measured vector line ar independence and the maximum space angle.