The life expectancy depends greatly on health - morbidity, disability and invalidity. A rise of life expectancy and increased morbidity make problems of the disabled people even more important; in many countries disability becomes more common. This paper presents methods of modeling the life expectancy without disability.The sample survey of disable persons was used to estimate the disability-free life table and to verify Gompertz, Weilbull, exponential and linear-expotential disability hazard functions. None of these functions was accepted to describe the disability hazard for the total surveyed population.The hypothesis that a probability of becoming disabled depends on age, sex, education, type of disability certificate, causes of disability, necessary support and subjective restraints from taking up a job was verified by the use of the semiparametric Cox model of proportional hazards. It was concluded that only age influence significantly the disability hazard. For people in the working age results of the disability-free life tables as well as Gompertz, Weilbull, exponential and linear-exponential disability hazard functions, estimated separately for the considered age groups, provided acceptable results except for the age 45-54. About 10 years before reaching the retirement age the disability hazard rate increased markedly and according to the estimated Cox model depends strongly on restraints in taking up jobs and needed care as well as on sex. Therefore disability hazard functions should be specified by sex.