<p>Quantum Machine Learning (QML) represents a frontier in computational intelligence, leveraging quantum mechanics to enhance data-driven learning paradigms. This study systematically evaluates the efficacy of two quantum machine learning algorithms-Quantum Support Vector Machine (QSVM) and Quantum Neural Networks (QNN) on IBM quantum simulation platforms. Utilising the Aer Simulator, Qasm Simulator and Statevector Simulator, the algorithms are benchmarked across six diverse datasets: Iris, Wine, Breast Cancer, Rain, Indian Liver Patient and Banknote Authentication. To accommodate the qubit limitations of current quantum systems, feature dimensionality is optimised through Principal Component Analysis (PCA) while preserving the majority of data variance. Performance is evaluated using accuracy, precision, recall, F1-score, training loss and execution time. In addition, classical machine learning models, including Support Vector Machine, Decision Tree, k-Nearest Neighbour and Multi-Layer Perceptron are employed as baseline classifiers using the same preprocessing pipeline and feature space. Experimental results demonstrate that the performance of quantum models is highly dataset-dependent, with competitive results achieved on selected complex datasets where the performance of classical models degrade while classical approaches generally exhibit superior accuracy and computational efficiency on low-dimensional benchmark problems. The findings provide a comprehensive assessment of QSVM and QNN under current Noisy Intermediate-Scale Quantum (NISQ) constraints and contribute to understanding the practical applicability and limitations of quantum machine learning in near-term quantum environments.</p>