The etiology of motion sickness is now usually explained in terms of a qualitatively formulated "sensory conflict" hypothesis. By consideration of the information processing task faced by the central nervous system in estimating body spatial orientation and in controlling active body movement using an "internal model" referenced control strategy, a mathematical model for sensory conflict generation is developed. The model incorporates and extends models proposed by von Holst, Held, and Reason, and is congruent with multisensory models for spatial orientation developed by Young and coworkers. The model postulates a major dynamic functional role for sensory conflict signals in movement control, as well as in sensory-motor adaptation. It accounts for the role of active movement in creating motion sickness symptoms in some experimental circumstances, and in alleviating them in others. The relationship between motion sickness produced by "sensory rearrangement" and that resulting from external motion disturbances is explicitly defined. A nonlinear conflict averaging model is proposed which describes dynamic aspects of experimentally observed subjective discomfort sensation, and suggests resulting behaviours. The model admits several possibilities for adaptive mechanisms which do not involve internal model updating. Further systematic efforts to experimentally refine and validate the model are indicated.