Steam condensation plays a key role in removing heat from the atmosphere of the Westinghouse AP600 containment in case of a postulated accident. A model of steam condensation on containment surfaces under anticipated accident conditions is presented and validated against an extensive and sound database. Based on the diffusion layer theory and on the usage of the heat/mass transfer analogy, one can deal with large temperature gradients across the gaseous boundary layer under high mass flux circumstances. The thermal resistance of the condensate film, as well as its wavy structure, have also been considered in this model. As compared to Anderson's experimental database, an average error lower than 15\%, within the experimental confidence range, has demonstrated its remarkable accuracy. In particular, the model has shown a good response to the influence of primary variables in steam condensation (i.e., subcooling, noncondensable concentration and pressure), providing a mechanistic explanation for effects such as the presence of light noncondensable gas (i.e., helium as a simulant for hydrogen)in the gaseous mixture. In addition, the model has been contrasted against correlations used in safety analysis (i.e., Uchida, Tagami, Kataoka, etc.) and occasionally to Dehbi's database. This cross-comparison has pointed out several shortcomings in the use of these correlations and has extended the model validation to other databases.