This paper addresses the issue of error accumulation in Laser Inertial Simultaneous Localization and Mapping (SLAM) for large-scale environments in intelligent mobile devices by proposing a global constraint-based LiDAR SLAM algorithm for global environment map construction and localization. The algorithm consists of two parts: In the part of GNSS/IMU loose combination and LiDAR SLAM, 18 state parameters are estimated and solved by Kalman filtering method in the GNSS/IMU loose combination, and the GNSS/IMU combination positioning results are added as global constraints on the basis of traditional LiDAR/inertial SLAM. The cumulative errors of laser SLAM can be effectively reduced, the accuracy of positioning and mapping of mobile devices in large-scale and outdoor environment scenarios can be improved, and the global positioning information can be provided to realize seamless indoor and outdoor positioning. The results show that the errors in X, Y and Z directions decrease from 1.538 m, 0.345 m and 28.56 m to 0.536 m, 0.115 m and 0.851 m, respectively, after adding GNSS/IMU global constraints. The improvement in map construction and positioning accuracy, as well as the enhanced accuracy of point cloud mapping with repeated scans, confirms the effectiveness of the proposed algorithm.