Air pollution from motor vehicles in densely populated cities remains an urgent problem. In order to analyse and predict air quality at city scale, in-situ monitoring systems of pollutant concentrations and meteorological parameters are used, supplemented by mathematical prediction models and numerical simulations. However, local pollution processes in urban areas located close to roads often cannot be studied efficiently by monitoring due to the lack of a detailed sensor network. Numerical modelling at the urban micro-scale can provide detailed information on the pollutant transport around buildings, but requires high resolution building geometry and approaches to describe boundary and initial conditions for meteorological parameters, pollution sources and traffic intensity.In the present study, the authors focus on the numerical modelling of scenarios of pollutant transport from vehicles in the vicinity of an urban environment fragment in Stuttgart, Germany. The modelling methodology is based on a combination of classical CFD methods, sensor data on traffic intensity and vehicle statistics, and meteorological data. The traffic statistics used in the simulation include the number of vehicles per hour, divided into 11 vehicle classes, and vehicle speed data. The numerical simulation was performed using the 3D Reynolds-averaged Navier-Stokes equations with the k-ε turbulence model, supplemented by gaseous pollutant and particulate matter transport models. Ansys Fluent 2023R1 is used as the main CFD modelling tool.Non-stationary measured data of traffic on the considered road section were analysed and used to improve the numerical model. This allows the description of the daily air pollution dynamics within the urban fragment. This detailed simulation of unsteady aerodynamics and air quality at the urban micro-scale can be used to focus on indoor/outdoor air exchange problems, to optimise the transport system, and to contribute to the provision of sustainability and climate safety technologies in cities.