In longitudinal observational studies, longitudinal variables are often correlated with observation times. Also, there may exist a dependent terminal event that stops the follow-up. In this article, we propose a joint modeling approach for analyzing longitudinal data with informative observation times and a terminal event. This approach introduces a shared frailty to specify the dependence structure among the longitudinal process, the observation and terminal event times. Some estimation procedures are developed for the model parameters and the degree of dependence. The asymptotic properties of the proposed estimators are established. The finite sample performance of the proposed estimators is examined through simulation studies. An application to a medical cost study for chronic heart failure patients from the University of Virginia Health System is provided.