Lung cancer is a common cancer that reveals notable death statistics worldwide. Due to the uncontrollable spreading rate of abnormal cells in the lung, lung cancer has become the leading cause of cancer death, with an approximate estimation of 1.8 million. In this context, early diagnosis has been highlighted to increase the survival rate. To address this challenge, this study evaluated two machine learning models, namely SVM and XGBoost for predicting risk levels. The models were trained and evaluated using a lung cancer dataset which consists of 1000 samples with 26 clinical parameters. The dataset underwent pre-processing steps and exploratory data analysis to efficiently handle missing values, outliers, and feature selection. After pre-processing, the dataset was split into training and test sets. Then, the predictive models were trained by using the training set, and the performances were assessed through the test set. Hyperparameter tuning was applied to optimize the model performance and avoid overfitting. Both SVM and XGBoost give promising results for all performance parameters assessed including accuracy, precision, recall, F1-score, and ROC-AUC. These findings demonstrate the potential of machine learning in early lung cancer detection.