Abstract: A script from different website forms a service mashup and its main characteristics are aggregating the data and combining them to make a visualizable outputs. A web service is used to exchange data between application and systems with a collection of open protocols. Through KDS (Knowledge Discovery in Services) on an open source of services from the internet we visually represent mashup to understand and predict the information in more accurate precision. In KDS according to the use we classify the matching technique from local to global, various complementary domains, merging more than one to same domain. The matching technique is classified synthetically into three kinds of layer they are granularity, kind of input layer, techniques layer. Here, comes the matching technique falls under element level matching technique which analyses entities or instances in isolation also ignoring their relations with other entities or their instances. The input interpretation layer chooses the syntactic technique with regard to its sole structure. The kind of input layer given to it in the form of terminological where strings found in the ontology descriptions. String based technique under element level are followed with more similar string, the more likely they are to denote the same concepts. Using distance function maps a pair of strings to a real number. This study introduces a process for KDS to assess the ability to predict service mashups using a combination of inputs and outputs.