Objectives: This study aims investigate the results of sputum cytology, which began its test period from 2009 as a special health examination for individuals whose lungs have been exposed to carcinogens. Methods: 22,215 cases of sputum results were classified into six categories through Papanicolau classification. The sputum results were analyzed according to age, exposure degree to welding fumes and smoking status. Chi-square tests and logistic regression analysis were used to compare and analyze the associations between sputum results and smoking status, exposure degree and age classifications. Results: The number of class Ⅰ cases was 20,784 with a frequency of (93.6%), class Ⅱ had 1,076 cases (4.8%), class Ⅲ had 69 cases (0.3%), and the inadequate class had 286 cases (1.3%) and there were no cases in class Ⅳ and Ⅴ. Patients in class Ⅲ received CT scans of their chest for further evaluation, but no cancer patients were found from that. There were 2.48 times (95% CI 1.12~5.53) more smokers than non-smokers in class Ⅲ and 1.70 times (95% CI 1.43~2.02) more in class Ⅱ. The number of ex-smokers was 2.46 times (95% CI 1.01~5.99) larger than the number of non-smokers in class Ⅲ and 1.28 times (95% CI 1.03~1.60) larger than the number of non-smokers in class Ⅱ. There were 1.54 times (95% CI 1.28~1.85) more patients from the mid-exposure group than the low exposure group in class Ⅱ. There were 1.62 times (95% CI 1.31~1.99) more patients from the high-exposure group than the low-exposure group in class Ⅱ. Conclusions: As a screening test for lung cancer, the usefulness of sputum cytology is limited. However, the precancerous cells (class Ⅱ, class Ⅲ) varied according to welding fume exposure degree and smoking status.