We propose a lattice model of two populations, predators and prey. The model is solved via Monte Carlo simulations. Each species moves randomly on the lattice and can live only a certain time without eating. The lattice cells are either grass (eaten by prey) or tree (giving cover for prey). Each animal has a reserve of food that is increased by eating (prey or grass) and decreased after each Monte Carlo step. To breed, a pair of animals must be adjacent and have a certain minimum of food supply. The number of offspring produced depends on the number of available empty sites. We show that such a predator-prey system may finally reach one of the following three steady states: coexisting, with predators and prey; pure prey; or an empty one, in which both populations become extinct. We demonstrate that the probability of arriving at one of the above states depends on the initial densities of the prey and predator populations, the amount of cover, and the way it is spatially distributed.