A lane-changing behavior is an important component of traffic simulation. Lane-changing actions are normally confined to the decision-making process of the task, and the action itself is assumed as an instantaneous event. Besides, the lane-changing behavior is based mostly on the observable positions and speeds of other vehicles, rather than on their intentions. As a matter of fact, changing one lane requires an average of about five to six seconds to complete. Existing lane-changing models do not comprehensively consider drivers’ reactions to work zone lane-changing signs (and other related messages, if any) and lane-changing actions. Furthermore, drivers’ socio-demographics are normally not carefully considered in conventional lane-changing models. With regard to this, the fuzzy logic-based lane-changing models that consider drivers’ socio-demographics are developed to improve the realism of lane-changing manoeuvres in work zones. Drivers’ Smart Advisory System (DSAS) messages are provided as one of the scenarios. Drivers’ socio-demographic factors are primary independent variables, while Lane-Changing Action Time (LCAT) and Distance (LCAD) are defined as output variables. The model validation process yields acceptable error ranges. To illustrate how these models can be used in traffic simulation, the LCAT and LCAD in work zones are estimated for five geo-locations with different socio-demographic specifications. Results show that the DSAS is able to instruct all drivers to prepare and change lanes earlier, which thereby shortens the duration of changing lanes. Educational background and age are essential variables in the developed models, whereas the impacts of gender on the output variables are indistinctive.