We propose a site-occupying model (SOM) to evaluate hydrogen solubility in metal because there is a discrepancy between the concept of the site-blocking model and experimental fact, as shown in this work. SOM is based on the experimental fact that hydrogen atoms in metal cannot come closer to each other than about 0.21 nm under hydrogen pressure of 1 MPa or less. We evaluate the hydrogen solubility by developing a statistical-mechanical method which considers partition functions for translational motion, configuration, vibration, rotation, and spin weights. SOM is verified by agreement with experimental data on hydrogen solubility in metal. We also compare SOM with the site-blocking model, and find that the theoretical curves from SOM agree well with the experimental data even at high hydrogen solubility of over 0.5 H/M by atomic ratio (hydrogen/metal).