We present and demonstrate a method based on a periodic phase noise estimation in an optical frequency domain reflectometry (OFDR) capable of a beyond 107 range-resolution-1 product (RRP) for the first time, which corresponds to 2.5 × improvement compared with the state-of-the-art. The moving average filter is employed to suppress the amplification of noise in the derivation process. Further, with the help of a third-order Taylor expansion, this method provides a highly precise estimation of periodic phase noise, which is the main factor impacting the performance of OFDR systems with medium-to-long measurement range combined with a submillimeter spatial resolution. A spatial resolution within 535 μm over the measurement range of 8 km is obtained. The proposed method offers a promising technique for fiber network monitoring and sensing applications.