Aiming at the problem that accurate positioning cannot be achieved by using only lidar in airport test environment, a tightly coupled multi-sensor simultaneous localization and mapping (SLAM) method based on factor graph optimization is proposed. The edge feature points and the surface feature points are combined with the two-step optimization strategy to match the point cloud features, and then the sliding window is used to eliminate the outliers of the global navigation satellite system (GNSS) data, and the GNSS factor is added incrementally when the point cloud is sparse and the robot azimuth changes more than 20°. The LeGO-LOAM, LIO-SAM and this method are compared and verified in the airport auxiliary road and the relatively empty airport runway. The results show that the absolute error of the trajectory of this method in the airport auxiliary road environment is 15.426 and 0.211 m lower than that of LeGO-LOAM and LIO-SAM respectively. After adding the GNSS factor, the accuracy is improved to centimeter level (RMSEAPE =0.050 m); in the airport runway environment, the trajectory of this method is relatively smooth, without offset problem, and the accuracy reaches centimeter level (RMSEAPE=0.063 m).