<p>To enable automatic reasoning about the original relations of 3D basic rectangular cardinal direction relations, this work aims to provide spatial data analysis with the capability for intelligent reasoning and predicting direction relations. High-performance search for original relations of rectangular cardinal direction relations is essential for problems such as composition, inversion, and consistency checking of 3D basic cardinal direction relations. This work presents an algorithm for the original relations of 3D basic rectangular cardinal direction relations. The algorithm searches and prunes the solution space with the help of a queue, enabling automated reasoning and computation of the original relations of 3D rectangular cardinal direction relations under the direction relation matrix model. Theoretical analysis demonstrates the correctness and logical completeness of the algorithm. Experiments show that in rectangular cardinal direction relations with conventional non-zero element counts (6, 8, 9, 12), the algorithm achieves an average efficiency improvement of 33% through the pruning strategy. In extreme cases with high non-zero element counts (18, 27), computational efficiency drops sharply due to the exponential growth in the number of original relations. The algorithm effectively addresses the problem of computation with the original relations for 3D basic rectangular cardinal direction relations, laying the foundation for the automated reasoning and analysis of issues such as the composition, inversion, and consistency checking with 3D basic cardinal direction relations, and then enhances and improves the intelligent analysis and processing capabilities of spatial databases for spatial direction relations.</p>