Image is considered as one of the most popular forms of information. Users of social media make extensive use of images to convey information. The invention of low-cost smartphones and digital camera has led to the creation of huge repository of images. This has made the task of easy access of images quite challenging. In order to easily access relevant images from the database containing millions of images, image retrieval systems play an important role. Image retrieval systems make use of either text or visual features to search for relevant images. Text-based retrieval systems make use of keywords to retrieve images. However, they require tagging of large number of images with relevant text in order to retrieve correct images. Content-based image retrieval (CBIR) system utilizes visual features of image to retrieve visually similar images. CBIR techniques either use low-level features or high-level features of image to retrieve visually similar images. While low-level features represent visual characteristics of image efficiently, they fail to provide semantic information present in the image. The high-level features overcome this limitation by extracting semantic features from the image. This chapter discusses the concept of intelligent techniques for CBIR which utilize high-level features of image to construct feature vector. The use of intelligent techniques not only helps in bridging the semantic gap but also achieves high retrieval accuracy as compared to low-level feature-based techniques. This fact has been demonstrated with the help of the proposed method based on intelligent technique. It utilizes low-level features along with an intelligent technique to construct feature vector for retrieval. The experimental results prove the effectiveness of the proposed method.