The mining time complexity is more with the conventional MapReduce algorithm particularly with large data sets or big data. The existing MapReduce algorithms will take time and data sets have to be disclosed for analysis. Defining data inherently with fuzziness will reduce the retrieval time and actual data sets need not be disclosed. In this paper, Fuzzy data mining methods are discussed. With these methods the original data sets need not be disclosed for analysis. The fuzzy functional dependency (FFD) and fuzzy multivalve dependency (FMVD) are studied for data mining. The fuzzy MapReduce algorithm of FFD and MVFD are useful to reduce time of OLAP and the original data need not be disclosed, The FFD and FMVD are made easy for combining the data sets and used for association, some examples are given for fuzzy MapReduce.