Abstract: The existing risk decision-making models can only explain the selection results of some individuals. The fundamental reasons are attributed to three aspects: the emotion of an individual is not included; the single decision pattern and one-way data fitting method are unreasonable. The following four measures have been adopted to solve these problems. Emotion functions are designed for the outcome and its likelihood respectively, according to its general characteristics. The additive relation between emotion functions is determined, based on the controlled information processing model. A multiple decision pattern is used instead of the single decision pattern. Finally, a two-way fitting method is established to replace the one-way fitting one. Therefore, a risky decision-making model containing emotion is established to explain all individuals' choices. The model is checked and verified by the classic cases of Prospect Theory and 14 anomalies. The developed model provides a novel solution to the problem of risky decision-making.