With the development of autonomous driving and communication technology, the heterogeneous traffic flow by combining connected and autonomous vehicles (CAVs) and manually driven vehicles (MVs) will appear on the freeway in the near future. It is expected that CAVs can improve the freeway capacity and reduce vehicle exhaust emissions, but sharing the same road by CAVs and MVs will cause certain interference to CAVs. In order to reduce the negative influence of the heterogeneous traffic flow, setting up CAV dedicated lanes to separate CAVs from MVs to a certain extent is regarded as a reasonable solution. Based on the characteristics that MVs should be decelerated by a realistic amplitude and that the connected and autonomous vehicle can accurately predict the speed of its preceding and rear CAVs at the next time step, a heterogeneous traffic flow model was established. Based on this model, we studied the overall influence of different lane strategies on the operating efficiency of freeway traffic flow and vehicle exhaust emissions under different densities with different CAV penetration rates. The results show that setting up CAV dedicated lanes with low CAV penetration rates will have a negative impact on the freeway traffic flow. When the CAV penetration rate is 40%–60% and the density is not less than 30 veh/km/lane, setting up one CAV dedicated lane is the best choice. When the CAV penetration rate exceeds 60% and the density is not less than 40 veh/km/lane, setting up two CAV dedicated lanes is the best choice. The research finding will assist in understanding the overall influence of CAV dedicated lanes on freeway traffic flow and help determine the optimal number of CAV dedicated lanes under different traffic conditions.