Fatigue is an important issue for orthotropic steel deck (OSD) bridges and therefore a design criterium. Cracks will initiate at several locations of the bridge deck. It is likely that a crack at the trough web to deck plate joint at the crossbeam will initiate first. This crack initiates at the weld root and propagates through the deck plate. A crack at this location is harmful because it is not optically visible and therefore hard to detect. The first part of the report focuses on the fatigue assessment of the trough web to deck plate joint at the crossbeam. Fatigue assessment of this joint is possible with different assessment methods which are the nominal stress method and the hot spot stress method as given in the Eurocode, and the effective notch stress method, described by the IIW. A hand calculation using a simple 2D model is made and a finite element (FE) model is used for the assessment. Both a shell and solid element model are build. The fatigue life calculation of the joint for a deck segment is made. Experimental results for the same deck segment are available and used for validation of the FE models. The fatigue life of each method is compared to the determined load cycles to failure from the experiment. There is concluded that the hot spot stress method results in the most accurate fatigue life prediction and a solid element model gives a slightly longer fatigue life than using shell elements. The effective notch stress method results in a conservative fatigue life for this specific joint. Stresses in the zone near the weld root of the rib-to-deck plate joint at the crossbeam are in compression, while fatigue cracks will initiate due to tensile stresses. During welding, residual stresses appear in the joint. Therefore, a fatigue life prediction is made which includes these residual stresses. An approximation of the residual stresses is included in the initial state of the finite element model. Using the Smith, Watson and Topper (swt) parameter, which is a local critical plane based fatigue assessment, the amount of load cycles to failure is determined. The result is comparable to the experimental result. The second part of the report is focused on the geometry of the deck segment. A shell element model and the hot spot stress method are used to analyse different parameters. When the width between the trough webs is larger, a thicker deck plate thickness is required. Four alternatives for a bridge deck are made. The alternatives have a different amount of troughs over the width, which means that the width between the trough webs differs per alternative. The deck plate thickness is chosen so that each alternative results in a damage value just below 1 for traffic category 2 and a design life of 50 years. Fatigue load model 4 is considered. The weight of the alternatives is determined. There can be concluded that reducing the amount of troughs over the width results in a heavier deck. This is because the weight of the deck plate is a large part of the total weight of the OSD and if more troughs are used, a thicker deck plate is applied. There was assumed that the stress in the joint only is affected by deformation of the deck plate. However, the in-plane deformation of the crossbeam has a negative effect on the stress range in the trough web-to-deck plate joint. Besides the geometry, the effect of some assumptions for modeling are investigated. Including a load dispersal through the surface layer by applying a larger wheel contact area results in a reduced fatigue damage value of about 50% and is therefore beneficial to take into account in the design calculations. The stress range in the joint is highest if a wheel load is placed in the center of the trough. A transverse shift of the wheel results in a large decrease in stress. Therefore it is important to take into account the design load cycles per transverse wheel position.