<p>Machine learning focuses on making computer systems learn from data and make decisions based on acquired knowledge. Hidden Markov model (HMM) is a probabilistic model in machine learning for representing and processing sequential data. To enhance the performance of sequential analysis, we propose a quantum hidden Markov model (QHMM) that transforms each component of classical HMM into its quantum counterparts. The proposed system consists of an encoding method that converts emission probabilities of observation onto quantum states and utilizes Gram-Schmidt orthonormalization to convert transition matrices into quantum operators. Moreover, it employs mid circuit measurement to carry hidden-state information across time steps. To evaluate the performance of the proposed QHMM, we apply part-of-speech (POS) tagging on English corpora and the English-Hindi code-mixed dataset. A comparison between the performance of the proposed model and the classical HMM has also been shown, where QHMM provides higher F1-score compared to classical HMM.</p>