The purpose of this research is to develop a suite of privatization algorithms for symbolic systems, specifically systems modeled as Markov chains and Markov decision processes, to protect the sensitive user data used to generate these models. We use differential privacy as our framework to provide privacy protections to user data. To enforce differential privacy for symbolic systems, we develop algorithms for privatizing reward functions, model constraints, and transition dynamics. Additionally, we propose work to privatize the trajectories generated by these models, to develop a filtering framework for these privatized trajectories, and to develop quantitative rules that automate algorithm selection. Such rules will enable users to assess which algorithm best fits their privacy needs, and which algorithm to use if more than one can meet these needs.