Lung cancer is one of the biggest concerns in the medical industry today, and high accuracy survival predictions are crucial for making informed decisions in patient care. In our research project, we applied various classification techniques to a dataset containing symptoms of patients that was taken from Kaggle to classify lung cancer patients in terms of survival. We focused on comparing the predictive power of Logistic regression (LR), Decision Trees (DT), Gaussian Naive Bayes (GNB) model, Support Vector Classifier (SVC) model, K Nearest Neighbour (KNN) model, Random Forests (RF) model and Voting Classifier ensemble model. The key data attributes used in our analysis included age, smoking, chest pain, anxiety, and number of such symptoms that might lead to lung cancer. Our Voting Classifier ensemble, which combined the predictions of the individual models, demonstrated the best performance on taking the top 4 best performing classifier models together with the mean accuracy of 94.90% (Hard voting) and 94.05% (Soft Voting). Among the six individual models, DT and RF performed the best with a mean accuracy of 94.70%, followed by SVC with an accuracy of 93.58%. Decision Trees underperformed, likely due to the limited number of discrete outputs. Our Voting Classifier approach outperformed the individual models, highlighting the benefits of combining multiple models to improve predictive accuracy. The result also indicated that the SVC model generated a distinctive output, suggesting its potential utility in specific applications. Our findings indicate that ML models can provide comparable performance in predicting lung cancer patient survival. These results suggest that the application of techniques utilizing machine learning to lung cancer datasets can be useful for estimating patient survival time and informing patient care decisions. Future work includes expanding the dataset and exploring other machine learning techniques and ensemble methods to further improve the prediction performance.