This paper examines analytical models for hail impact on composite laminates and compares the predictions with finite element simulations and experiments. The crushing of the ice results in a distributed load and a much higher delamination threshold load than for impact by hard objects. Furthermore, prediction of the impact load by merely considering the mass flow of ice particles results in too low loads and a response in disagreement with experiments. The pressure acting on the impacted plate is fairly uniform within the contact area, but initial through-thickness waves during the first moments of the impact cause much higher stresses than the quasi-static values assumed in the analytical models. Hence the finite element models predict a different load history, with a much steeper initial increase in the contact load.