The increasing complexity and dynamic change of massive data processing currently receive widespread attention. One of its core content is to study how to use the existing information to achieve rapid updating of knowledge. Granular computing (GrC), a new computing paradigm of information processing, is an emerging research field which is mainly used to describe and deal with uncertain, vague, incomplete and massive data, and provides a solution based on the granularity and the relationship between the granularities. As an important part of GrC, rough set theory is an effective mathematical tool to deal with the uncertainty and imprecise problems. Based on the MapReduce model in cloud computing, this paper first presents a parallel algorithm for computing the equivalence classes, decision classes and the association between them in rough set theory. A parallel algorithm is then designed for computing rough set approximations from large-scale data. To adapt to the dynamic real-time system, the MapReduce model and incremental method are combined to build two parallel incremental algorithms for updating rough set approximations in different incremental strategies. An extensive experimental evaluation on big data sets show that the proposed algorithms are very effective and have better performance with the increasing size of the data.