One of the main environmental problems we are currently facing is air pollution. Air quality models calculate how much pollution is emitted and dispersed into the atmosphere. This research presents a Computational Fluid Dynamic model using a real urban geometry for the analysis of CO contamination with a three-dimensional model. This method includes a procedure of calculating emissions using different types of vehicles. CO Measurements are obtained from a Wireless Sensor Network to validate the models. The present study analyzes six representative real cases of different traffic situations and climatic conditions plus 3 hypothetical cases in a hotspot area in the city center of Valencia. The results show what influences pollution levels the most is the wind direction, which influences the generation of velocity patterns. In the validation cases, the real wind direction is used and a slight change produces great differences in both velocities and CO concentration. In the hypothetical cases, parallel and perpendicular winds are defined to observe the differences when this ideal situation is applied. In conclusion, the mixing and transport of air pollutants are closely related to the structures of velocity and turbulence that occur in the air, which depends strongly on the wind direction.