Abstract: In various wireless sensor networks applications, the sensor node positioning problem is the basis and key technology of IOF (internet of things), which is essential for IOF monitoring activities. Due to the influence of environment and many other factors, the data collected by sensor node is easily generate errors, leading to the existing location algorithm or model positioning effect not so satisfactory. Aiming to this problem, this paper using internet (Geometric) topology structure information, based on the model of Location Estimation-Locality Preserving Canonical Correlation Analysis (LE-LPCCA), through transforming the depiction of signal space and physical space local information, proposed a new IOF positioning model LE-RLPCCA. Compared with the kind methods at present, model LE-RLPCCA has a better robustness of error-tolerance technology, positioning accuracy and stability improved evidently.