Lung cancer is a worldwide threat to humanity since its cells grow uncontrollably inside lungs leading to increased mortality rate. The lung cancer often poses serious breathing issues and it is been contributed majorly by smoking and inhaling smoke. Even with high medical advancements, the effective treatment and curing of lung cancer are not effective till date. Proper precautions and earlier stage detection may reduce the cancer to spread the entire organ. It is hence necessary to check with least minimal data i.e. text data can provide even a greater effectiveness in diagnosing the lung condition. Meta-heuristic algorithm influences greatly with its computational capability and offers stronger prediction of lung cancer at earlier stage with accurate analysis. In this paper, we develop a classification system using flower pollination algorithm (FPA) that tends to classify the medical text documents. The FPA algorithm classifies the medical text documents to diagnose the lung cancer in humans. The FPA is applied as an intelligent algorithm that imitates the behavior of pollination in flowering plants to identify the essential classes of lung cancer. It finds the relationship between the pollens to identify the essential classes based on flower position. The simulation is conducted to validate the effectiveness of the model with other meta-heuristic optimization methods that include bee colony optimization, and ant colony optimization algorithm. The results of simulation show that the proposed method undergoes effective classes of lung cancer than other existing methods that includes accuracy, sensitivity, specificity, f-measure and mean average percentage error.