In this work, the effect of a distribution of dipolar interaction fields on the magnetization of superparamagnetic nano-granular system is examined using a general Monte-Carlo model. The model consists of a cubic cell in which the particles’ locations and directions of easy axes are generated randomly. The particle sizes are generated according to a lognormal distribution with a median diameter Dm = 6 nm. The room temperature magnetization curves at different particle packing densities are simulated. The calculations show that the magnetization curves are always depressed with increasing packing density. Usually, this effect would be taken as denoting predominantly demagnetizing interaction effects (i.e., negative dipolar fields). We have investigated the distribution of dipolar interaction fields along the x, y, and z directions and found them to be symmetric and Gaussian in form with a mean very close to zero. The net effect of the transverse field components is to reduce the magnetization. The same is true for the component in the direction of applied field, Hiz, despite the fact that positive and negative dipolar fields are found to be equally probable. According to this picture, it is not correct to say that dipolar interaction fields are negative. Thus, we ascribe the reduction in magnetization to the non-linear response of the magnetization to the applied field, which weighs the negative interaction fields more strongly than the positive fields. It is concluded that the non-linear response of the magnetization is equally important as the sign of the interaction field.