لاتین: owadays with due regards to the development of medical, economic and industrial studies in the field of survival and longitudinal data analysis, the use of advanced statistical techniques is required more than ever. In survival studies, because of censored data, special modeling techniques are used. The related models can be divided into two categories: the parametric category in which a statistical distribution is considered for the survival time, such as accelerated failure time models and the nonparametric category like Cox proportional hazards models that mainly model the hazard function instead of survival time. Usually in regression survival analysis, the hazard rate for different subjects, due to the effects of unobserved factors related to various subjects is different. In this regard, frailty models are taken into consideration by entering subject effects as random variables in the model. Usually with a collection of survival data, other variables are also registered over time. Such data that are obtained by repeatedly measuring variables over time for different subjects is called longitudinal data. In this type of data, usually serial correlation among observations over time is considered by entering lagged-response variables in the model. In this case, it is necessary that the dependence of initial response variables with the random effects associated with subjects’ effects is also considered. Ignoring this correlation may lead to the serious bias in the estimation of model parameters. Handling this issue is by modeling of initial responses and taking into account the dependence between this model and the original equation through assuming shared random effects. In many studies, the survival time depends on the longitudinal variables that are measured over time. Therefore, we need to the joint modeling of longitudinal and survival data. The dependence between longitudinal data and time to event data could be due to the influence of same subjects for longitudinal variables and the survival time. In such circumstances, the shared-parameter model is one of the most famous models that are used for analysis of data. On the other hand, the dependence of longitudinal and time to event data could be due to the direct influence of longitudinal data on the survival time that should be correctly addressed in the fitted model. In this thesis, while introducing different models of survival analysis and analysis of longitudinal data, common models for joint modeling of survival time and longitudinal data are introduced. Also, a joint model is proposed to cover the serial correlation of longitudinal observations and also their influences on survival times. Moreover, another model is proposed to address the direct influence of longitudinal data on survival times through entering the longitudinal variable as a time-varying covariate in the fitted model for the survival time. Because of the complexity of estimation of parameters in joint models, with the Bayesian approach to the problem, Markov chain Monte Carlo simulation methods are adopted. In this regards, full conditional posterior distributions required for running the Gi sampler are derived for the proposed models when working with the accelerated failure time models.