Logistic regression model was applied to study and analyze the effect of economic and social factors on job satisfaction as a nominal variable includes two groups (satisfied, dissatisfied on work). These factors represented by seven quantitative variables: the worker's age, number of experience years, monthly salary, the number of family members, the pressures of work, opportunities of advancement and development, and the work environment, as will three categories variables: the social situation of worker, academic qualification, and job title. Bayesian method was used as advanced method to estimate the parameters of the proposed model and application study relied on a random sample data of size 473 workers from workers of «Cairo West Power Station», Cairo, Egypt. Gibbs sampling method was applied by using Win Bugs program to generate a 51,000 sample from joint posterior distribution of logistics parameters, probability intervals obtained and showed that each of the pressures and worries of work, work environment, experience years, and educational qualification of the worker has a significant effect because probability interval for each of them does not contain zero value. Goodness of fit test proved that the logistic regression model of job satisfaction in a «Bayesian» enough to reconcile the data. Predictive ability of the model shown that the correct classification percentages exceeding 80% and the estimated model by Bayesian method appropriate to predict the status of worker (satisfied, not satisfied). The comparison between accuracy of the results related to each of maximum likelihood (ML( and Bayesian method using error rate, as well as the (ROC( analysis method was conducted. The results of application error rate method showed that no differences between ML and Bayesian method with middle sets of satisfaction probabilities, in the last sets Bayesian method more accurate than the ML method, but ML method more accurate than the Bayesian