This paper introduces a new inhomogeneous stochastic 1/2-power Lundqvist-Korf diffusion model to forecast the gross national income (GNI) per capita of Morocco. Through a transformation and the application of It’s calculus, we identify the probabilistic properties of the process, including its analytical form. We then derive the probability transition density function (ptdf) and the mean functions associated with the model. Discrete sampling and the maximum likelihood approach are used to estimate the model parameters. However, because of the complexity of the associated system of likelihood equations, we employ the simulated annealing procedure to deal with the estimation problem. To test the methodology, we first apply the model to simulated data and evaluate the statistical results. To finalize, the model is used to fit real data corresponding to Morocco’s GNI per capita.