A stochastic differential equation model for wireless fading channels is considered and the problem of estimating the model parameters from irregularly sampled data is studied. A computationally very efficient estimator based on the least squares method and the introduction of a differentiation operator is suggested. It is described how the operator should be chosen to obtain a consistent estimator as the number of data tends to infinity and the upper bound on the irregular sampling interval goes to zero. Attention is given to the computation of the operator. An online implementation of the estimator is also included.