This paper presents an inter-query semantic learning approach for image retrieval with relevance feedback. The proposed system combines the kernel biased discriminant analysis (KBDA) based low-level learning and semantic log file (SLF) based high-level learning to achieve high retrieval accuracy after the first iteration. User's relevance feedback is utilized for updating both low-level and highlevel features of the query image. Extensive experiments demonstrate our system outperforms three peer systems.