Object-Oriented System Testing (OOST) focus on issues emerged with Object-Oriented features e.g. encapsulation, polymorphism, inheritance and dynamic binding. Different faults can detect during the interfacing between classes: interface faults, conflicting functions, and missing functions. With iterative nature of testing process, Traditional (automation) testing techniques become less efficient in forecasting new defects resulting in pesticide paradox. To overcome the limitations of traditional inter-class testing techniques, automation testing techniques need to be powered by artificial intelligence for bring dynamic testing techniques into testing process. This paper presents a new dynamic approach to overcome the pesticide paradox in inter-class testing of object-oriented applications that stores the knowledge into ontologies and providing algorithms, which operate on the knowledge to regenerate testing steps easily with required modification to uncover defects. Hence, ontologies can be modified without changing the algorithms, and vice versa omitting using the same test cases to overcome the pesticide paradox. The proposed approach generates an executable test suite in five phases omitting using the same test cases to overcome the pesticide paradox. To validate the proposed approach, a tool entitled PSCCOTM (Polymorphism State Collaboration Class Ontology Test Model) is developed and a case study is applied using PSCCOTM tool. The results show that, test cases can be easily updated by uploading modified ontology file of a test model into PSCCOTM tool. Also, the execution results show high percentage of faults detection, however new cases studies need to be implemented to confirm the attained results.