An adaptive integrated Kalman filtering based on the adjustment outputs of local navigation sensors and the outputs of a dynamic or kinematic state model is presented, which avoids the correlations of the local Kalman filtering outputs affected by the same disturbances of the dynamic state model. It has the advantage of rigor in theory and simple in calculation as well as adaptive in the various local navigation outputs. An integrated navigation estimator that is similar to the federated Kalman filtering is given as an initial estimate of a state based on the information sharing principle, but without any dynamic model information. An adaptive integrated fusion of the local navigation outputs and the dynamic model information is followed, in which the weights of the local navigation outputs and the dynamic model outputs are determined based on their differences from the integrated navigation results. The processing algorithms, logic, and associated computer burden are similar to those of federated filter. A simulated example is given to show the effectiveness of the new adaptive integrated navigation algorithm.