Railway rail fastener system is prone to defects such as fastener missing, elastic strip fracture, elastic strip skew, and bolt loosening or over tightening. The traditional detection method based on two-dimensional intensity images can identify the first three diseases, but it is difficult to detect the bolt looseness or over tightening defect. In order to solve this problem, a detection method of fastener bolt fastening state based on line structured light is proposed. The 3D point cloud data of track structure are obtained by 3D camera. The rail area and fastener area are segmented by building a height integral function and prior knowledge, and the height difference between the bolt to be measured and the non-wear area outside the rail head is calculated. Based on this, a method to determine the threshold of bolt fastening state based on online updating threshold library is proposed. The height difference is compared with the threshold database to realize the detection of bolt fastening state. The experimental results show that the detection rate of fastener looseness or over tightening is more than 80% under the condition of detection speed of 20 km·h ^-1. This method makes up for the shortcomings of the traditional fastener detection methods and realizes the automatic detection of fastener fastening state.