Analytical methods that predict the endurance of structural wood members in a fire are based on the reduction of the cross section of the member caused by wood being charred. To define the charring rate in terms of more fundamental properties, empirical models were established. Eight species were tested for charring rates and material properties. Regression analysis was used to develop the models. The predictor variables for the initial factorial design included density, moisture content, treatability, and hardwood-softwood classification. The addition of char contraction simplified the model and reduced the predictor variables to the char contraction factor, density, and moisture content. Our results show the importance of surface recession and moisture content to wood charring.