Special vehicles (SVs) are vehicles which conduct tasks such as the maintenance of urban roads and are typically characterized by travelling at a lower speed at a constant rate of speed within the same lane. In order to reduce the influence of SVs, guidance zone is designed and provides traffic guidance suggestions (TGS) for human‐driven vehicles (HVs) helping drivers for better decision between car‐following (CF) and lane‐changing (LC). To verify the effectiveness of TGS, an improved Dogit‐agent‐based model is established to simulate the captive and not captive choice of CF and LC for different driver types under TGS, and build the rules for mixed traffic flow of SV and HVs. Finally, a numerical simulation with a three‐lane system is conducted to analyze the traffic efficiency through a set of indicators, and the results show that the TGS can reduce the influence of SVs on traffic flow in a specific occupancy rates range, increase the cross‐section traffic volume by about 5%. The TGS also can increase the average speed of HVs in the lane behind SV by about 5% to 30%, and increase traffic density to 200% on the underutilized lane in the raw space in front of the SV.