A generalized model for a big data warehouse based on the NoSQL technology is proposed. This technology does not impose strict requirements to a database structure. It is scalable in large extend with a high-speed processing of queries and also provides the ability of database multiple servers replication with automatic data synchronization support. Data warehousing with NoSQL technology allows the design of databases to be different from the traditional relational database management systems (DBMS) based on SQL query tools for data access. The MongoDB database system is used for satellite metadata management. The source of data for the developed system is satellite-based spectroradiometer MODIS data (Terra and Aqua platforms). Gdalinfo utility from GDAL library is used for satellite data files processing. A test analysis is performed for the warehouse of more than 20 Gigabytes. The dual-core processor server with a CPU clock speed of 2.1 GHz and RAM 2 GB is utilized. The time for completing the entire sequence of operations is estimated in about 5 minutes.