Life satisfaction (LS) is one of a set of constructs defining quality of life. Previous studies showed that LS was sometimes related to health and sometimes not. The study was designed to examine the relation of LS as a general construct to satisfaction in specific domains. We assumed that there is a tendency to maintain an acceptable level of LS even under stressful and threatening conditions, that it is related to optimism and that the likelihood of attaining satisfaction in a particular domain affected the selection of domains on which LS is based. We expected that in cancer patients LS would be related to more domains but not to health. The study was done with 55 head-and-neck cancer patients, of all stages and grades of tumor; 51 orthopedic patients, victims of accidents with good recovery chances; and 55 healthy individuals. The healthy individuals and orthopedic patients were matched (in terms of group values) to the cancer patients in age, gender and education. Single-item measures of LS and optimism, and a questionnaire with 49 multiple-choice items assessing adjustment in 13 domains were administered to all subjects. The results showed that in cancer patients LS was related to most domains but not to health and not to optimism, whereas in the other groups it was related to few domains including health, and also to optimism. The findings support the tendency to maintain LS with the materials available to the individual, and show that health is related to LS only if its maintenance or attainment are realistic goals. Thus, both bottom-up and top-down theories of LS are supported.