This paper addresses the problem of estimating the air gap for a large semisubmersible production platform. Although it has a great impact on the design of the floating unit, many times the minimum deck height is still defined from simplified methods that incorporate relatively large safety margins. The reason for this is the intrinsic complexity of the associated hydrodynamic problem. Nonlinear effects on the incoming and scattered waves are usually relevant and sometimes nonlinear effects on the motions of the floating hull may also play an important role. This discussion is illustrated by means of wave basin tests performed with the model of a large semisubmersible designed to operate in Campos Basin. Significant run-up effects on its squared-section columns were observed for the steepest waves in several design conditions. Also, the unit presented relatively large low-frequency motions in heave, roll and pitch, which also affected the dynamic air gap measurements. In order to evaluate the difficulties involved in modeling such phenomena, simplified tests were also performed with the model fixed and moored in regular waves of varying steepness. Wave elevation in different points was measured in these tests and compared to the predictions obtained from two different numerical methods: a BEM code that incorporates second order diffraction effects (WAMIT) and a VOF CFD code (ComFLOW), the latter employed for fixed model tests only. Results show that a standard linear analysis may lead to significant errors concerning the air gap evaluation. Extending the BEM model to second order clearly improve the results as the wave-steepness increases. Although the VOF analysis is considerably time-consuming, simulations presented very good agreement to the experimental results, even for the steepest waves tested.