The invention provides a big-data parallel processing system and method based on column storage and relates to the field of big-data processing. The big-data parallel processing system comprises a client-side, a master control node and multiple data nodes, wherein the master control node comprises an HBase master controller and an SQL main engine. The HBase master controller takes charge of management and maintenance of the data nodes, the SQL main engine takes charge of analysis of SQL statements and continuous distribution of the SQL statements to the data nodes. Each data node comprises an HBase subarea node and an SQL sub engine, wherein the HBase subarea node takes charge of data storage and management on each data node, and the SQL sub engine takes charge of analysis and execution of SQL on the data nodes. The master control node and the data nodes comprise HBase tables and further respectively comprise SQL tables. The big-data parallel processing system and method based on the column storage is suitable for an environment with large-scale data volume and provides an SQL interface of a perfect structuralized relational data model.