Video frame rate up-conversion (FRUC) is a crucial video processing technique designed to enhance temporal resolution, improve motion smoothness, and optimize the viewing experience. Traditional FRUC methods are mainly categorized into non-motion-estimation (NME) and motion-estimation-based (ME) interpolation algorithms. While NME approaches have low computational complexity, they often introduce ghosting and blurring artifacts. In contrast, ME-based methods offer higher-quality interpolated frames but suffer from high computational complexity, making real-time processing challenging. To achieve efficient and accurate frame interpolation, we propose a novel FRUC framework based on a self-similarity model, combined with an optimized block-matching motion estimation algorithm. The proposed method leverages multi-resolution motion estimation to reduce computational costs and employs an improved motion vector smoothing technique to mitigate ghosting and block artifacts, thereby enhancing the visual quality of interpolated frames. Experimental results demonstrate that our approach outperforms conventional FRUC methods in both visual quality and computational efficiency, making it a promising solution for high-definition television, video streaming, and other real-time video applications.