Abstract: Existed monitoring models of dam foundation horizontal displacement adopt only one key monitoring point'sdata, while the choice of the key point is subjectiv e. Traditional multi-sensors method is inclined to result in inconsistency between different monitoring points. Kalman filter fusion algorithm is adopted to eliminate the inconsistency and make full use of the multi-point monitoring information. Fusion algorithm calculation is according to centralized method and distributed method with and without feedback separately. Then, fusion result is presented, and feasibility is comparatively analyzed. Further, an example simulation is given in order to illustrate the effectiveness of the method in application. Meanwhile, an effective theoretical method is provided for the multi-point monitoring for the hydraulic structure foundation as well.