The determination of plasmaspheric electron density profiles has been attempted using a stochastic approach, using Omega signals observed on the Akebono satellite. Since the wave normal directions and delay times of Omega signals can be theoretically calculated by ray tracing under an appropriate electron density model, the density profile can be estimated by fitting the calculated directions and times to the observed values. In the present paper, we introduce a flexible model and a novel algorithm in the fitting method, taking into account the stochastic factors of the density distribution and wave propagation. The stochastic representations of directions and times enable us to separately estimate the effects of the ionosphere and the plasmasphere. The validity of the proposed method is examined using observational data collected during the recovery phase of a magnetic storm. In another example, an estimation of the asymmetry of the plasmasphere is attempted.