Data migration of an application from source to target information system is a critical step for a successful modernization project. There are few commercial tools available to address the data migration challenge, creation of a data transformation specification is largely a manual, knowledge intensive, and expert driven process. A system and method for learning based synthesis of data transformation rules have been provided. The system is focused on automating important aspects of automatic inference of the transformation specification. The key principles behind the system and method are derived from the observations on how experts use domain, system, and historical mapping knowledge while creating data transformation specifications. The system contains two major components, schema matching and transformation rule program generation. The system uses machine learning, knowledge representation for schema matching and developed rule generator using a deductive synthesizer.