In unknown or confined spaces，autonomous navigation of ROS-based intelligent robots presents challenges such as low map accuracy and information loss，hindering real-time navigation. This paper introduces a design focused on a multi-sensor fusion-based autonomous navigation ROS intelligent robot. It collects environmental data using both LIDAR and depthcameras， employing a fusion of LIDAR-based SLAM and RGB-D visual algorithms to achieve multi-sensor fusion， consequently accomplishing simultaneous localization and mapping(SLAM) indoors for robot autonomous navigation.Experimental results demonstrate that the multi-sensor fusion mapping offers high localization accuracy and rich information.The constructed maps exhibit high correspondence with the actual environment，enhancing the robot's environmental perception and navigation flexibility，enabling the system to autonomously navigate accurately.