Machine learning in current world presents itself across various domains where data plays major role. It involves process from data collection to decision-making. This paper mainly focuses on the application of Hidden Markov Models in machine learning, aiming on investigating forward and Viterbi algorithms. Forward algorithm focuses on determining the probability of an observed sequence. On the other hand, Viterbi algorithm predicts the most likely sequence of hidden states based on the observed sequence. With the help of practical examples, this research demonstrates how HMMs can model scenarios where future outcomes are influenced by hidden factors. Beyond this, the potential of these algorithms extends to broader fields, including artificial intelligence, where these algorithms can be used to predict and control the unnatural behavior of artificial intelligence based systems by finding the hidden factors leading to unimaginable outcomes. As AI continues to grow across various domains or industries, the ability to predict, assess, and control such behaviors using HMMs becomes increasingly valuable. This research outlines the advantages, limitations, and future directions for HMMs in real world machine learning applications.