Folksonomy allows user to tag resources independent of vocabulary and is being widely used in social web systems. But these vocabularies have some limitations like ambiguity, variation between tags and many more. Tag based ontologies are developed to overcome such issues. This paper focuses on eliminating limitations of folksonomies and extracting semantic relationship between tags to develop ontologies. Finally semantically related tag is recommendation to user as per the user search query based on developed ontology. Also the user is recommended with the latest trending tags used in social web.