The fusion of object detection and relative positioning has emerged as a pivotal yet challenging domain with the increasing integration of smartphones into unmanned systems. This paper introduces an innovative algorithm, SUS-ODRP, designed to seamlessly combine object detection and relative positioning within smartphone unmanned systems. Firstly, SUS-ODRP leverages an improved You Only Look Once v5 (YOLOv5) framework as the cornerstone for object detection training. Secondly, SUS-ODRP reconstructs the heterogeneous binocular cameras located on the rear of a smartphone into standard binocular cameras. Finally, utilizing the object detection outcomes, SUS-ODRP calculates the relative coordinates and distances of the objects from binocular images. Experiments on a real smartphone unmanned vehicle demonstrate the efficacy of the SUS-ODRP algorithm, showcasing high detection precision, real time performance, and a high average positioning precision of 95.33%.