The load on the highway bridge includes dead load, live load, dynamic load, environmental load, special load (breaking and collision forces), etc. where traffic load is the predominant variable action. Traffic varies from site to site and with time; the uncertainty associated with traffic action is relatively high compared to other variable actions (Bailey 1996). Design traffic load models can be therefore be conservative in many cases because they are based on the most aggressive traffic to be found in the region of application of the given design code. There is a great potential for reducing traffic action models by considering actual traffic during the evaluation of an existing bridge (O’Connor and O’Brien 1999, Bailey 1996). The bridge chosen for this paper is a two-lane structure in the city of Vienna. The reliability of the structure was previously assessed using deterministic loading models. It is argued that a better indication of the structural reliability would be obtained using real traffic data from the site. SiWIM installed their equipment on the structure to provide site-specific, continuously recorded, data on truck weights, speeds, time of arrival, headway, axle configuration, volume, etc. The paper presents the prediction of characteristic extreme load effects to which the bridge may be subjected using the measured traffic flow data from SiWIM and theoretically generated Monte Carlo (MC) traffic files. The accuracy of the later is directly proportional to the amount and quality of measured information available (O’Connor and O’Brien 1999). Statistical extrapolation of the results of simulations, under the assumption of stationarity, permits determination of the characteristic load effects (Jacob 1991) for the chosen structure. 2 Equipment installation