In Taiwan, motorcycles are the major means of transportation used for short distance trips. The number of registered motorcycles has already risen to 15 million. Among the victims in traffic crashes, the percentage of motorcycle users (including motorcyclists and passengers) exceeds 80%. Motorcycle accidents incur a great social cost. In an effort to improve motorcycle rider safety and traffic flow efficiency, numerous roadway design features and management strategies have been devised for motorcycles (with engine capacity of 250 cc or less). For example, motorcycles may be prohibited in the inboard lane closest to the oncoming traffic, and motorcycles may be required to make a two-stage left turn at intersections where there are three or more lanes. However, current safety design features can only partially deal with conflicts between motorcycles and other vehicles, and several traffic conflict problems in straight lanes are still unsolved; for example, motorcycle riders going straight have conflictswith taxi drivers and bus drivers going to the roadside to pick up passengers. Moreover, the width of lanes on a roadway is designed from the view of car drivers. Risky behavior by motorcyclists, such as weaving in traffic and speeding, increases the accident risk for motorcycle riders. These issues highlight the need for improved roadway design for motorcycle safety in situations of mixed traffic flow. This study aims to develop engineering methods of roadway design for motorcycle safety in situations of mixed traffic flow with the consideration of motorcycle rider behaviors and factors contributing to crashes. This study will analyze the road crash database built by the National Police Agency, Ministry of the Interior. The following factors will be considered to identify the critical accident types for motorcycle accidents: types of collisions, types of roads, separation designs for lanes in different directions on a road, separation designs for lanes in the same direction, traffic control facilities (e.g., traffic signals), and traffic conditions (e.g., traffic flow). Based on the results of the macro analysis mentioned above, this study will sample certain patterns of accidents and collect more detailed information about motorcycle rider behaviors, other vehicle driver behaviors, road situations, traffic conditions and safety problems using data recorded by video cameras on roads, in vehicles and on helmets worn by motorcycle riders. The detailed video data will help us to micro-investigate and understand the causes of accidents. Finally, based on the safety problems and traffic engineering issues found in the macro and micro analyses, this study will provide different roadway designs for motorcycle traffic safety in situations of mixed traffic flow. This study will also evaluate the risks presented by the different designs.