Smoking rates in the military are evaluated through questionnaire surveying. Because the accurate identification of smokers facilitates the provision of smoking cessation services, this study conducted urine cotinine concentration testing to verify the accuracy of self-reported smoking behavior by female volunteer soldiers and analyzed the effects of second-hand smoking on urine cotinine concentrations. This study is a cross-sectional study conducted using purposive sampling on female volunteer soldiers receiving training at the Taichung Recruit Training Center in May 2014. This study simultaneously collected questionnaires and urine samples, and urine samples were analyzed with an enzyme-linked immunosorbent assay. The self-reported smoking rate of female volunteer soldiers was 19.3%, whereas the smoking rate as determined by urine cotinine concentration testing was 26.3%, indicating an overall underestimation of 7.0%. Chi-square (χ2) goodness of fit test results indicated that the distribution of self-reported smoking behaviors and that verified from urine cotinine concentration testing were significantly different. The sensitivity of self-reported smoking behavior was 66.7% with a specificity of 97.6%. There was no significant association between second-hand smoking and urine cotinine concentrations. Questionnaire survey self-reporting methods could underestimate the smoking behavior of female volunteer soldiers and routine testing with biochemical verification is necessary.