This paper presents Marker-Assisted Region Scanning for Simultaneous Localization and Mapping (MARS-SLAM), a novel approach to optimizing the Simultaneous Localization and Mapping process in unknown environments. The method was specifically designed to address the challenges of autonomous exploration in extreme conditions, to enable efficient navigation and offer a systematic approach to determine the completion of mapping. The approach uses markers to indicate unexplored regions, ensuring an organized and complete exploration. During the process, the robot places markers in free areas identified by the LiDAR sensor, located at the sensor’s range limit, building a list of regions yet to be explored. The mapping is considered complete when the marker list is empty, indicating that no unexplored regions remain. Target marker selection during navigation is based on age and distance. Age refers to the chronological order in which markers are created, while distance refers to the length of the route from the robot to the marker. The method is validated in two virtual environments of varying complexity. Experimental results demonstrate the effectiveness of MARS-SLAM in achieving complete mapping and accurately identifying mapping completion. Compared to alternative navigation methods, including predefined zigzag routes and routes generated by Ant Colony Optimization (ACO) algorithms, MARS-SLAM shows superior performance, particularly in reducing the number of poses required to complete the mapping, achieving a 64.39% reduction in poses compared to ACO and a 71.07% reduction compared to zigzag.