The concept of the quasar main sequence is very attractive since it stresses correlations between various parameters and implies the underlying simplicity. In the optical plane defined by the width of the Hβ line and the ratio of the equivalent width of the Fe II to Hβ observed objects form a characteristic pattern. In this paper, we use a physically motivated model to explain the distribution of quasars in the optical plane. Continuum is modelled as an accretion disk with a hard X-ray power law uniquely tight to the disk at the basis of observational scaling, and the Broad Line Region distance is determined also from observational scaling. We perform the computations of the FeII and Hβ line production with the code CLOUDY. We have only six free parameters for an individual source: maximum temperature of the accretion disk, Eddington ratio, cloud density, cloud column density, microturbulence, and iron abundance, and only the last four remain as global parameters in our modelling of the whole sequence. Our theoretically computed points cover well the optical plane part populated with the observed quasars, particularly if we allow for super-Solar abundance of heavy elements. Explanation of the exceptionally strong Fe II emitter requires a stronger contribution from the dark sides of the clouds. Analyzing the way how our model covers the optical plane we conclude that there is no single simple driver behind the sequence, as neither the Eddington ratio nor broad band spectrum shape plays the dominant role. Also, the role of the viewing angle in providing the dispersion of the quasar main sequence is apparently not as strong as expected.