Accuracy improvement is one of the significantproblemrepresented in the recommender systems in the research community. Currently, Multi-criteria recommender system used as a multi-criteria rating to estimate overall rating. In this paper, we represent types of recommendation system and primaryform of a multi-criteria recommender system in improving accurate prediction. And also describe some soft-computing techniques andmetrics involved in Multi-criteria recommendation system. Finally compared various techniques and the method used to find accuracy prediction in recommendation system. © 2018, Institute of Advanced Scientific Research, Inc. All rights reserved.