Information fusion is an important topic in scientific research. Soft likelihood function is a common method of fusing evidence from multiple sources. However, when the combined evidence contains equally important decision information, the fusion results obtained using existing methods do not reflect the attitudinal characteristics of decision makers. To address this problem, a novel generalised soft likelihood function is developed in this paper. First, a new notion of decision maker (DM) pair is defined, which is used to char- acterise the outcome of the decision as well as the reliability of the evidence. Then, a series of algorithms for correcting the initial evidence set data are formulated. Eventually, a generic soft likelihood function for fusing com- patible evidence information is proposed. Numerical examples are used to illustrate the effectiveness of the proposed methodology.