Acoustic mapping of ocean currents with Distributed Networked Underwater Sensors (DNUS) system is an effective and energy-saving technique for monitoring oceanographic environments. Unlike the common acoustic tomography methods, this new approach can locally reconstruct ocean currents using data between neighboring sensors. Considering that ocean currents are highly correlated during a short time interval, this paper develops a Kalman filter based tracking approach to improve the DNUS-based ocean current estimation. A 2-D ocean model is used to generate synthetic observational data. The simulation results show that, with the information of the previous ocean current estimates introduced, the proposed method can outperform the traditional current mapping method with DNUS-type system.