Abstract: Automatic image annotation (AIA) is a technique to provide semantic image retrieval. In AIA, the image contents are automatically labelled with a pre-defined set of keywords which are exploited to represent the image semantics. This paper proposes an Automatic image annotation method using MPEG-7 features. The feature vectors are provided for training the KNN-classifier. When a query image is provided by the user, the extracted features are provided to the trained KNN -classifier for annotation. The system also compares the results of different combinations of mpeg 7 descriptors. The Automatic image annotation proposed in this paper is on fruit images. This system has several applications like automatic labeling and price computing of fruits and vegetables in a grocery store, morphological analysis of fruits, for scientific studies, enhanced learning for kids and Down syndrome patients using fruits pattern recognition.