This chapter covers the topic of learning causal models from observational data. A general introduction to causal discovery is presented, highlighting some of the challenges. Then, the types of graphs used to represent partial models are introduced. Several algorithms for causal discovery from observational data are explained, including score-based and constraint-based methods, causal discovery with functional and parametric constraints, and techniques based on continuous optimization. To conclude the chapter, we address subject-specific causal discovery and a method to direct undirected edges based on causal effects estimation.