The interest in increasing the use of timber in structural engineering has risen significantly inrecent years due to the urgency of reducing the environmental impact of the building industry.Timber bridges is one of the many possible use-cases for timber in structural engineering. Whendesigning timber bridges the serviceability criteria regulate the design with deck accelerationsoften being the most limiting factor. It is therefore essential to accurately model the dynamiccharacteristics of timber bridges during the design stage. In the past this has proven difficultand more research on this area is needed and this theses aims to shed some light on this issue.An existing pedestrian timber bridge located in Brumundadal, Norway, has been analyzedin this master thesis. The aim was to extract the bridge’s dynamic characteristics using experimentalanalysis and to replicate that behavior in a numerical model while investigating theeffects of different modelling strategies.The extracted mode shapes from the experimental data are consistent for the first threemodes, all below 5 Hz. The damping for the modes below 5 Hz are above 1 %, the lowest being1.2 %.Both a simple and a detailed numerical model were created. It was found that the connections,masses and asphalt are vital factors in the dynamic behavior of the numerical models.The simple and detailed model complied with the experimental data similarly with the simplemodel having more accurate eigenfrequencies while the detailed model had more accurate modeshapes and impact responses. The good match of the simple model implies that combining certaincounteracting simplifications can produce a good model. However, without experimentaldata for calibration it may be difficult to know what simplifications results in an accurate model.To obtain models that match the experimental results two rotational stiffnesses were required.One for the pinned supports of the trusses and one for the connection between the trusses andthe cross beams. The rotational stiffnesses needed are perpendicular to the axis of the dowelsin the slot in steel plates. The stiffness of these connections could not be accurately predicteddespite efforts with advanced FE-models and analytical approaches. The stiffnesses for theseconnections that were used in the bridge models were instead determined by trial and error to getresults that match the experimental ones. An analytical model should be investigated for the rotationalstiffness along the perpendicular axis of the connector of slot in steel plates connections.Some design recommendations for timber bridges are: One should treat the asphalt as deadweight. Additionally for every boundary condition and connection in the model that has uncertainstiffnesses in one or more DOFs one should try modelling that DOF as either free or fullyrigid. One should then choose the case that produces the most conservative outcome. Boundaryconditions and connections should transfer DOFs in the centre of the joint. Steel plateconnections have a high weight compared to the timber which affects the dynamic behaviorand their mass should therefore be modeled accurately. The mass of the railings should alsobe considered. Finally it is highly recommended that the timber properties are defined as anorthotropic material instead of a isotropic definition. This is necessary even for beam elementsto accurately capture the timbers shear modulus. Note that the Poisson’s ratio barely affectsthe model behavior.