For the multisensor descriptor systems with correlated noises, the WMF full-order Kalman smoother is presented. First, the centralized fusion descriptor system is obtained. Then this centralized fusion descriptor system is transformed to the descriptor system with uncorrelated noises. Second, the weighted measurement fusion (WMF) descriptor system is presented. Finally, the WMF full-order Kalman and smoother and its smoothering error variance are presented, based on the maximum likelihood (ML) estimation and extended state method. A simulation example of a 3-sensor descriptor systems with common distribute noises verifies its effectiveness and the accuracy relation of the presented WMF full-order Kalman smoother.