---

---

In this paper, we first define longitudinal-dynamic heteroscedastic hierarchical  normal  models. These models can be used to fit longitudinal data in which the dependency structure is constructed through a dynamic model rather than observations. We discuss different methods for estimating the hyper-parameters. Then the corresponding estimates for the hyper-parameter that causes the association in the model will be presented. The comparison among various  empirical estimators  is illustrated through a simulation study. Finally, we apply our methods to a  real dataset.