Objective: To assess the reliability and validity of randomized response technique(RRT) model of cluster sampling survey on multichotomous sensitive question investigation method and statistical formulae. Methods:A total of 404 female sex workers were selected randomly by cluster sampling in Xichang during May 2011 to July 2011. RRT was applied to obtain their sensitive characters. With the sensitive characters taken as parameter of simulated overall,simulated direct questioning results taken as criterion,results got from multichotomous RRT model were compared with ones from simulated direct questioning by using chi-square test with Monte Carlo simulation to access validity and reliability. Results:There were 99 results have no statistically significant different from criterion (α=0.05). Conclusion:The method and corresponding formulae have high validity and reliability,and can be used for sample survey of large-scale sensitive issues.