The popularity of eco-friendly bicycling has shown a growing interest. As a means of transportation, bicycles interact with vehicles and pedestrians and are allowed to travel through facilities such as bicycle lanes, shared streets, and urban roads. Bicyclists are vulnerable to serious injury as they bear the brunt in car-to-bicycle crashes. Past studies have focused on bicycle left-turn passage methods in traffic scenarios involving manual vehicles (MVs), whereas studies on scenarios where autonomous vehicles (AVs) and MVs are mixed are insufficient. Therefore, the primary objective of this study is to present an optimal left-turn passage method by evaluating their safety, operation, and environment following the market penetration ratio (MPR) of AVs. The bike simulator data were collected and calibrated with actual driving data to reflect bicycle driving behaviour in the simulation. The passage methods for bicyclists analyse the hook-turn, bike box with priority signals, bike box without priority signals, and narrow bicycle lanes with bicycle signals. The impact of MPR was identified through countermeasures for each scenario. The results of this study could contribute to the preparation of greenhouse gas reduction and carbon-neutral policies.