Machine learning is a valuable tool for data analysis, but traditional models face limitations in classical computing, particularly with high-dimensional or complex data. Quantum computing offers potential advantages by leveraging principles like superposition, entanglement, and quantum parallelism to process information in ways classical systems cannot. This study focuses the potential of quantum-enhanced machine learning by comparing quantum long short-term memory (QLSTM) models with classical long short-term memory (LSTM) models. LSTM is widely used for time series prediction and sequence modeling because it can capture the time dependence of data. QLSTM is a hybrid quantum-classical model that incorporates quantum circuits to exploit quantum phenomena and has the potential to extend the capabilities of conventional LSTMs. The performance of QLSTM was evaluated in terms of learning convergence, test loss, and overall accuracy through experiments on real-world data sets. The first results showed that QLSTM outperforms conventional LSTMs, with faster convergence during training and lower test loss. These results suggest that QLSTMs have the potential to revolutionize machine learning by providing a more powerful tool for analyzing complex data with greater accuracy and reliability. However, more research is needed to fully realize the potential of data analysis with quantum functions, including optimizing quantum circuit design, managing computational overhead, and extending QLSTM models to large data sets.