An Intelligent Tutoring Systems is a type of knowledge-based system whose main agenda is to efficiently supplement a human tutor with a machine. Dissimilar to conventional classroom teaching, Intelligent Tutoring Systems (ITSs) have the ability to fit according to the necessity of an individual learner. More emphasis has been laid on various types of e-learning systems. In this work, a probability-based ITSs system is proposed consisting of four models specifically the learner’s behaviour model, pedagogical model, knowledge base model and learner assessment model. The importance has been given to the learner assessment model where an element of uncertainty has been introduced and handled by the Bayesian Belief Network (BBN). The purpose of the learner assessment model is to rightly detect the knowledge level of each learner based on their reply to the level of questions, where the level of questions is random to the process of assessment to the learner. The uncertainty factor has been defined in terms of success and failure parameters. Success is the probability that a learner of low cleverness level gives a right reply to a level of questions and is increased by a small probability of 0.07, whereas Failure is the probability that a learner of high cleverness level gives a wrong reply to a level of questions and is reduced by a small probability of 0.04. In this work during an assessment of the knowledge level of a learner, the system has incorporated the uncertainty factors of Success and Failure with the help of Bayes’ rule and has found promising results that take into account the possibility of Success or Failure.