Education and training is one of the vital energetic forces and an essential condition for a nation’s economic, social and cultural development. Education plays such a role as it rises and supports the creative of new things and productive capability of human beings. The overall objective of the new national TVET policy and strategy is to generate a skilled, interested, adaptable and advanced workforce that plays pivotal roles in the poverty reduction and socioeconomic development efforts of the country. The assessment center gives national assessment mainly for TVET trainee as assessment center report shows half of the trainees have failed due to unknown reason. So as take proactive action knows factors of trainee’s failure. The aim of this study is to design a knowledge-based system by integrating machine learning results with a knowledge base system, to determine the performance of the trainee in the training and assessment process. In this study experimental research methodology was followed, to conduct an extensive experiment is used to acquire knowledge automatically from the data set and represent it in the knowledge base. In this study, a knowledge-based system is planned for determining the performance of trainees. The knowledge base system was used a classification algorithm, specifically the Random forest by utilizing TVET trainee’s dataset to extract hidden knowledge from TVET trainee’s dataset with a performance evaluation result of 97% accuracy. The integrator makes connection of model which was created by the random forest classifier to a knowledge-based system to add knowledge automatically. Then the integrator understands the syntax of the random forest classifier and PROLOG and converts from random forest rule representation to PROLOG understandable format. To do this, java programming was used to integrate the results with the knowledge-based System. The proposed system of the study has registered a good performance. Hence, the performance of a prototype system in this study registered 86.6% and gain 88.45% of user acceptance. The major challenge of the developed system was when the number of rules increased in the knowledge base. The researcher recommends that integrating Android and SWI-Prolog will be better research for future researchers xii Keywords: Technical and Vocational Education and Training, Competency level, Trainee, machine learning, Knowledge Base, Automatic Knowledge Acquisition