According to the World Health Organization (WHO, Ambient (outdoor) air pollution (who), 2020), outdoor air pollution is estimated to have caused 4.2 million premature deaths worldwide in 2019. Under this scenario, the development of statistical tools for assessing the impact of a possible reduction of fossil fuels-based transportation systems on air quality is more relevant than ever. This type of instruments can help policy makers to take the right path to tackle the problem of air pollution in urban environments. Thus, this work proposes a method to evaluate the impact of the reduction of vehicle fleets on air quality. The method is based on the construction of counterfactual time series, which represent observed air pollutants under a treatment of interest. In this case, the treatment of interest are the vehicle fleets on the streets. The construction of the counterfactual time series is based on Bayesian dynamic linear models. Several impact assessment measures are implemented, taking advantage of the counterfactual and observed time series (the observed pollutant without the treatment). The flexibility of the Bayesian approach allows easy inference of the impact quantities via high posterior density intervals. Some simulation analyses show that the method works well under different impact scenarios, and one application regarding Medellin’s case in Colombia is presented. For the application, it is shown that the drastic reduction in vehicle fleets during the COVID-19 mobility restrictions led to a significant reduction in the presence of different pollutants ($$\hbox {PM}_{{10}}$$
PM
10
, $$\hbox {PM}_{2.5}$$
PM
2.5
, $$\hbox {NO}_{{2}}$$
NO
2
, $$\hbox {NO}_{{x}}$$
NO
x
)