RNA molecules perform numerous important biological functions including sensing small molecules, regulating gene expression, and catalyzing reactions in all living systems. For many of these processes, RNAs bind to proteins to form RNA-protein assemblies (RNPs) that carry out processes such as translation of RNA to proteins and splicing of pre-mRNA. The structures and energetics of these RNAs and RNA-protein complexes dictate their sophisticated behavior, but often elude rich experimental characterization. Predictive models of these features would enable powerful interpretation of sparse experimental data and efficient design of new RNAs and RNPs. My work aims to create the first computational tools to calculate RNA-protein binding landscapes, build three-dimensional structures of RNA-protein complexes, automatically model RNA coordinates into cryo-electron microscopy (cryo-EM) density maps, and to accelerate reliable RNA structure determination. My colleagues and I have rigorously assessed the predictive power of these methods through blind tests on functionally and structurally diverse RNAs and RNA-protein complexes. This work has culminated in the determination of several novel three-dimensional structures of RNA molecules including ribozymes, riboswitches, and computationally designed molecules. These results suggest that predictive models of RNA and RNP structure and energetics can now enable accelerated, unbiased structure determination when combined with rapidly acquired experimental data, and will be useful for understanding the complex biological functions of RNAs and RNA-protein complexes.