The aim of this paper is to estimate models for labor supply probabilityof household members by using Bayesian procedures. Our modelsdescribe the labor supply behavior of multiple household members. Particularly,we deal with the behavior of households with one, two or threenon-principal earners (a wife and/or children). The related literaturemainly studies the labor supply of one non-principal earner. The modelsare of discrete choice or qualitative response type to yield a laborsupply probability that non-principal earners accept an employment opportunity.We apply Bayesian estimation methods to the models and carry outempirical tests of the models using Japanese data.Douglas law of labor supply shows that the labor participation rateof wives are negatively correlated with husband's income. This law hasbeen reconfirmed by empirical analyses of household-level surveys.Although the analyses are limited to a case of one non-principal earner,it may be possible to say that the labor supply of multiple non-principalearners is also negatively correlated with husband's income.The situation is, however, complicated when we consider householdswith multiple household members. In fact, it is difficult to define laboruparticipation rate at a household level when the household has multiplenon-principal earners. We may define the rate as a ratio of number ofparticipants to that of non-principal earners within a household. We needa theoretical basis for the definition. We need a model to describe thelabor supply behavior of households with multiple non-principal earners.For this purpose we provide a general discrete choice model. Three modelsfor labor supply are derived from it.In applications of Bayesian methods, a difficult problem is how toformulate prior distributions. The diffuse or uniform prior distributionis usually employed. In this study we take up Bayesian methods for estimatingone model to incorporate the information obtained in the priorestimation of other models. We also use maximum likelihood methods forcomparison.We make some experimental calculations based on the estimated parametersto clarify theoretical implications of the models. We calculateelasticities of labor supply probabilities with respect to changes ofexogenous conditions.