Smoke exposure data among U.S. wildland firefighters for carbon monoxide, respirable particulate and respirable crystalline silica are presented from a field surveillance program between 2009 and 2012. Models to predict fireline-average exposure to each inhalation hazard were developed and fit to the available data. The models identify factors to consider when defining similar exposure groups and designing future data collection. Task-based rather than shift-average data collection is important because the work activity representing the majority of fireline time, the position up- or downwind of the fire, and the proportion of time this combination represented were significant factors in the model for carbon monoxide, and all but wind position were significant for respirable particulate matter. The wind position versus the fire was not important for respirable quartz exposure. The crew type was an important factor in each model.