The objective of my thesis is to enhance the understanding of compressive membrane action (arching action) in concrete slabs with special relevance to composite bridge deck slabs. Compressive membrane action is a common structural phenomenon in laterally restrained concrete slabs and enhances the loading capacity of laterally restrained slabs due to in-plane thrust derived from the restraint offered by the boundary conditions. The aim of this research is to establish the behaviour of bridge deck slabs using the nonlinear finite element method and experimental tests. More and more bridges built in the past 50 years employed composite structures with decks constructed of reinforced concrete and supported by longitudinal steel girders. A series of third scale composite steel-concrete bridge deck models were built for the experimental models. The design parameters ofthe concrete strength, the steel beam size and reinforcement percentages were varied in the models. The target of the· experimental tests was to find out the influence of these design parameters had on compressive membrane action. Commercial finite element packages were adopted to simulate the compressive membrane action in the concrete slabs using nonlinear numerical analysis. The accuracy of simulation results. Because punching failure is a common failure mode in this structural type and is difficult to be simulated in FEA, failure criteria based on implicit and explicit analysis were established to capture the limit loads for the numerical analysis. Furthermore, the research proposed the most suitable simulation method to be used and modelled compressive membrane action in concrete slabs, including the type of element and solution method to be adopted. At last, a modified theoretical model based on QUB model (Taylor et al 2003) for the prediction of ultimate loads of bridge decks was proposed.