Using sunlight and semiconductors to split water in a photoelectrochemical (PEC) process to generate hydrogen provides a viable way to produce renewable fuel. So far, the efficient PEC devices are exclusively based on epitaxial grown III-V semiconductors. However, these semiconductors are expensive to make, limiting their practical application for large scale PEC hydrogen production. PEC devices based on polycrystalline thin films can potentially lead to low-cost production of hydrogen by PEC. However, most PEC devices made of polycrystalline thin films have not achieved high efficiency yet. Searching for new semiconductors that can potentially lead to efficient PEC devices is still necessary.  In this talk, we will present general strategies for searching for such kind of semiconductors. We will show by density-functional theory that some new chalcogenides and oxides are promising candidates for efficient PEC applications. We will also present our latest results on experimental assessments on these new PEC materials.