Measurement bias and lack of terrain features often cause false peaks during underwater terrain matching positioning, that is, there is more than one peak near the real position. Previous methods to address this problem have increased the number of measurement beams, but this also increases the data processing time and energy consumption. At the same time, the ratio of measured information that is used does not increase. In other words, we should increase the ratio of measured information that is used, not simply increase the amount of information that is measured. Conventional matching algorithms only use the height of nodes without considering surface information, which is composed of height and the position of multiple nodes in three-dimensional space. Multi-beam sonar can obtain the three-dimensional distribution of terrain nodes. This node information is not just a height sequence, as it is used in previous methods. If we consider the nodes as a three-dimensional distribution of points with height and position information, this increases the matching position information and more of the terrain features can be extracted from the same measured data. Hence, in this paper, a terrain positioning method called the Node Multi-information Fusion (NMIF) is presented. This method focuses on improving the stability and accuracy degraded by bias in the Digital Elevation Map (DEM), terrain repeatability, and other factors. First, the concept of a Single Node Data Packet (SNDP) is introduced. The SNDP includes elevation and surface information surrounding the node, such as roughness, gradient, and slope. This additional topographic feature information improves the robustness and accuracy of the system. A computer simulation using actual ocean bottom topography verifies the advantages of the proposed NMIF algorithm.