Exploring unknown, GPS-denied environments poses critical challenges, especially in space missions where autonomous navigation is essential. Simultaneous Localization and Mapping (SLAM) enables Unmanned Ground Vehicles (UGVs) to navigate autonomously in such conditions, but reliable localization and mapping must overcome sensor noise, environmental uncertainty, and drift. This work presents a GPS-free, multi-sensor SLAM system for space UGVs that integrates data from a Light Detection and Ranging (LiDAR) sensor, a Monocular Camera (MC), and an Inertial Measurement Unit (IMU) through enhanced sensor fusion. A modified Graph-Based (GB) method fuses MC and IMU localization estimates to reduce drift, while an Extended Kalman Filter (EKF) processes LiDAR-based localization in a second fusion stage, refining the trajectory. Simulations conducted in a virtual environment validate the proposed method under realistic conditions. The results show significant reductions in positional drift, improved localization robustness, and superior performance compared to a standard odometry-based EKF. By combining complementary sensors in a modular framework, the approach improves SLAM for autonomous navigation in unstructured terrains and demonstrates strong potential for planetary exploration and other GPS-denied missions.