Options are financial contracts that function as insurance against unwanted price movements. The global turnover has, since the beginning in the 1970s, increased enormously and is currently estimated to be at least 70 000 billion US dollar annually. This number can be compared to the US GNP, which was approximately 12 000 billion US dollar last year, i.e. the turnover in option markets is several times larger than the US GNP. The earliest (and still used) models for valuating options did not work very well during extreme events such as the stock market crash in 1987 (Black Monday). Researchers have since that event focused on finding more general models, adding additional stochastic terms corresponding to changes in the market. Common for these improved models is that they start by postulating that the market dynamics is given by a fixed model and thereafter calculate the value of the option for that given model. This project takes a different approach on improving option valuation models by taking parameter and model uncertainty into account. Financial data will never provide a certain answer which "model" generated data, although it is often possible to rule out insufficient models. Agents can either choose a single model or look at different (a mixture of) models. We study the effects of using a mixture of parameters and/or option valuation models and use this to derive more accurate valuation methods.