Degree-days are important parameters for developing predictive models of mosquito populations. Conventional methods of estimation of degree-days are typically applied to observations under constant temperatures in the laboratory. These methods are difficult to apply in the field where temperatures fluctuate. For estimating degree-days of larval developments of mosquitoes in the field, we applied Dennis' stochastic phenological model to a semi-field observation of Culex restuans. A parametric bootstrap approach was introduced to estimate standard errors of the parameters. We found that the estimated degree-days varied with various assumptions about the temperature base. Given the latter, the model can satisfactorily estimate the degree-days of larval and pupal developments of Cx. resturans. This approach is especially useful to measure spatio-temporal variability in larval development among various types of aquatic habitats.