The stochastic Arctic sea ice model described as a single periodic non-autonomous stochastic ordinary differential equation (ODE) is useful in explaining the seasonal variability of Arctic sea ice. However, to be nearer to realistic approximations we consider the inclusion of long-term forcing implying the effect of slowly-varying ocean or atmospheric low-frequencies. In this research, we rely on the equivalent Fokker-Planck equation instead of the stochastic ODE owing to the advantages of the Fokker-Planck equation in dealing with higher moments calculations. We include simple long-term forcing into the Fokker-Planck equation and then seek approximate stochastic solutions. The formalism based on the Fokker-Planck equation with a singular perturbation method is flexible with regard to accommodating further complexity that arises due to the inclusion of long-term forcing. These solutions are then applied to the stochastic Arctic sea ice model with long-term forcing. Strong seasonality in the Arctic sea ice model combined with long-term forcing, changes the seasonal variability depending on the phase of the long-term forcing. The change includes the shift of mean and the increase or decrease of variance and skewness. Stochastic realisations show that the change of the statistical moments due to long-term forcing is realised by unusual fluctuations particularly concentrated at a specific time of a year.