Dozens of white dwarfs with anomalous metal polluted atmospheres are currently known to host dust and gas discs. The line profiles of the Ca II triplet emitted by the gas discs show a significant asymmetry. In recent decades, researchers have also discovered several minor planets orbiting such white dwarf stars. The most challenging burden of modelling gas discs around metal polluted white dwarfs is to simultaneously explain the asymmetry and metal pollution of the star's atmosphere over a certain period of time. Furthermore, models should also be consistent with other aspects of the observations, such as the morphology of the emission lines. This paper aims to construct a self-consistent model to explain the simultaneous white dwarf pollution and Ca II line asymmetry over at least three years. In our model, an asteroid disintegrates in an eccentric orbit, periodically entering below the star's Roche limit. The debris resulting from the disintegration sublimates at a temperature of 1500 K, producing gas that viscously spreads to form a disc. The evolution of the disc is studied over a period of 1.2 years (over 21000 orbits) using two-dimensional hydrodynamic simulations. Synthetic Ca II line profiles are calculated using the surface mass density and velocity distributions provided by the simulations, taking into account for the first time the asymmetric velocity distribution in the disc. An asteroid disintegrating on an eccentric orbit gives rise to the formation of an asymmetric disc and asymmetric Ca II triplet emission. Our model can explain the periodic reversal of the redshifted and blueshifted peak of the Ca II lines caused by the precession of the disc on timescales of 10.6 to 177.4 days. Our work suggests that the persistence of Ca II asymmetry over decades and its periodic change in the peaks can be explained by asteroids on eccentric orbits in two scenarios. In the first case, the asteroid disrupts on a short timescale (a couple of orbits), and the gas has a low viscosity range ($0.001< to maintain the Ca II signal for decades. In the other scenario, the asteroid disrupts on a timescale of a year, and the viscosity of the gas is required to be high, $