In this paper we propose a soft computing-based solution to assess multivariate prediction ability in legal decision-making process. The proposed method is capable of incorporating prior knowledge. The main goal is, automatizing legal decision making, which is complex, complicated and often characterized by broad vigilance. The aim is helping legal practitioners to be able to focus not only on the decision itself, rather the correlations between them and combining of their variations as well. This way a "hybrid" thinking would be performed. The system will be able to "think through" a great number of possible variations of decisions and their combinations synchronized with the speed of the human mind.