Abstract Models of labor supply derived from stochastic utility representations and discretized sets of feasible hours of work have gained popularity because they are more practical than the standard approaches based on marginal calculus. In this paper we argue that practicality is not the only feature that can be addressed by means of stochastic choice theory. This theory also offers a powerful framework for developing a more realistic model for labor supply choices, founded on individuals having preferences over jobs and facing restrictions on the choice of jobs and hours of work. We discuss and clarify how this modeling framework deviates from both the conventional discrete approach [Van Soest, A. (1995) Structural models of family labor supply. A discrete choice approach. Journal of Human Resources 30: 63–88), as well as the standard textbook approach based on marginal calculus (Hausman, J.A. (1985) The econometrics of nonlinear budget sets. Econometrica 53: 1255–1282]. We furthermore discuss how the model based on job choice can be applied to simulate effects of alternative restrictions on hours of work.