Abstract: The purpose of data fusion is to merge or synthesize two data sets so as to make them become a better and complete data set. Based on extended inter-polation method, generalized fusion method is a method which strives to create a mathematical model that consists of two or more sets of equation and data sets. This method is not only able to bring out its full potential of data and measurement resources, but also capable of using innovative integration and mutual algorithm to solve the data fasion problem. After simulation and real measurement data analysis, conclusion can be drawn that compared with the least square algorithm, the generalized fusion method is more suitable for the repair of the data generated in the process of data fusion.This paper also introduces the main steps of the establishment and the solution of the generalized mathematical model, analyzes the distribution principle of weight in the model and its application in the field of satellite navigation data processing. Experimental results show that the method is practical and stable, and it is expected to be a new method for solving the problem of data fusion.