The mobility of vehicles on a selected segment of the transport network of an urban agglomeration is one of the four main parameters of transport flow. The article proposes an approach to estimating this parameter based on visual observation data. As an example, a section of Heydar Aliyev Avenue in Baku (Azerbaijan), characterized by the intensity of vehicle traffic, was selected. The information base of the study was made up of sensor readings from the Technical Vision System of the Intelligent Transport Management Centre of the Ministry of Internal Affairs of the Republic of Azerbaijan, which recorded vehicle speeds at the exit of this section every 10 s. The analysis and assessment of transport mobility was carried out in the paradigm of the fuzzy time series forecasting, reflecting the daily dynamics of changes in the interval of vehicles passing a selected segment of the urban transport network. The choice of the fuzzy paradigm is dictated by considerations relative to the weakly structured of averaged data from visual observation of vehicle speeds.