Abstract-Cancer is a rapidly spreading lethal illness that accounts for 10% of all deaths in Bangladesh. There are over ahundred different types of cancer. Predicting cancer is critical to the advancement of data mining applications. In thisstudy, the naive Bayes, k-nearest neighbour, and j48 algorithms are applied to predict cancer illness. Naive Bayes is asimple algorithm that is extremely beneficial when dealing with large datasets. K-nearest neighbour takes a dataset anddivides it into multiple groups, as well as predicting the categorization of fresh points. The J48 Classifier is built on thedecision Tree from training datasets, and it takes use of the fact that each of them and data sets may be utilised to decideon the smallest subset.The Weka tool is used to assess the correctness of a cancer illness dataset that includes nine differentcancer types. Cancer illness is predicted using 10-fold cross-validation. The accuracy of Naive Bayes is 98.2%, k-nearestneighbour accuracy is 98.8%, and j48 accuracy is 98.5 percent.Keywords- Data Mining, Lung Cancer Prediction, Classification, Naive Bayes, Bayesian Network, J48.