The big data technology provides a new data storage solution and computing model for processing massive seismic observation data. In order to solve the problems of low read/write speed,low concurrency performance and poor scalability of existing storage schemes based on relational database,we take earthquake precursor observation data as an example,and analysis of the business requirements in detail,and then propose the two earthquake big data storage solutions based on HBase and OpenTSDB. We also construct a test platform to compare the query,insertion and concurrency performance under different storage schemes. The experimental results show that HBase and OpenTSDB-based storage schemes have better scalability and concurrency performance than relational database storage scheme,and HBase-based storage schemes have the best read and storage performance.