For the practical issue of multisensor data fusion in target tracking, this paper combines the predicted values and the measured values weighted together. The optimal weight allocation method of Containing predicted values and measured values is presented. This method is equivalent to adding a sensor, which theoretically improves the accuracy of the fusion. At the same time, the maneuvering frequency function is established based on the firstorder difference vector of adjoining times and the current statistical model is improved. This change can improve the tracking effect. The measured results show that the maneuvering frequency function of this paper has a good performance on the tracking targets and the method combining with the measured values and predicted values is superior to the optimal weight allocation method of standard.