A method for estimating the uncertainity in the prediction of carbon dioxide concentration in a ventilated airspace is described. Carbon dioxide concentration is sensitive to measured data such as outdoor carbon dioxide concentration, indoor source, and ventilation rate which are stochastic in nature. This method is based on Ito stochastic differential equation which provides the statistical characteristic of the state variables. The method can simultaneously consider randomness in the input and the coefficients. Randomness in this case is modelled as a Gaussian white noise process. The moment equations are developed which provide the mean, variance and skewness of carbon dioxide concentartion. This information is useful in design of the ventilation systems. The verification and practical use of the model are presented.