Abstract: In this dissertation, the problem of distributed fault detection and data fusion for dynamic systems is investigated especially in the context of integrated vehicle health management. The common dynamic process is monitored by multiple redundant sensors. Normal and faulty behaviors can be modeled as two hypotheses. Due to communication constraints, it is assumed that sensors can only send binary data to the fusion center. There is a complicated correlation structure among sensor observations, and distributed detection with dependent observations is not tractable. Under the independence assumption and employing two correlation models, efficient and effective distributed fault detection algorithms are proposed, including local detector design and decision fusion rule design. These are based on state estimation via parallel distributed particle filtering. The independence assumption simplifies the design significantly and works well for a wide range of possible correlations among sensors. However, if the correlation is greater than a certain threshold, the design under the independence assumption fails in that using multiple sensors may perform worse than using just a single sensor. Examples show that better performance can be achieved by taking into account the correlation among sensor observations by employing the two correlation models than the one under the independence assumption. In addition to the hypothesis testing formulation, we consider distributed change detection for dynamic systems by joint system mode and state tracking. A distributed particle filtering approach in conjunction with the use of histograms to approximate local posterior distributions is proposed which reduces communication requirement and computation complexity. By increasing the number of sensors, the uncertainty associated with the system mode reduces, which leads to better detection performance. Simulation results show the efficiency of the proposed algorithm. The performance is not very sensitive to the number of bins used for approximation.