Abstract: Empirical studies as questionnaire and models are combined. Risk preference parameters in prospect theory is used to describe the characteristics of the insured risk preference, insured are divided into three categories. In accordance with the characteristics of the insured risk preference by cluster analysis: loss risk rationality type, sensitive to loss probability type and sensitive to loss value type. Characteristics of the insured risk preference are analyzed based on statistical data analysis. Combined with performance-based insurance contracts, insurance contract framework consist of base premium and incentive premiums has been designed for different types of irrational risk preference insured. As for sensitive to loss probability type insured, they can get incentive premiums by setting deductibles. As for sensitive to loss value type insured, full insurance model change into co-insurance when loss beyond the agreed full insurance range. Finally, this type insurance contract can stimulate disaster prevention efforts from the insured, so as to achieve the objective of reducing the risk and reducing the probability of loss.

Key words: irrational risk preference, prospect theory, cluster analysis