The purpose of this papaer is to investigate the risk caused by roadside parking vehicles. The data used in this study were observed by VTR camera at a four-lane arterial street. At the observed section, roadside parking is prohibited, but some vehicles usually park at the roadside in front of a shopping store because of the insufficient parking capacity of the store. Swerving and lane-changing pattern of straight through vehicles caused by roadside parking vehicles are classified and the accepted gaps of swerving and lane-changing vehicles are analyzed by use of observed data. A probabilistic model is built to estimate the risk of accident based on the headway distribution and the result of the gap acceptance analysis.