Based on a popular MJO monitoring method designed by Wheeler and Hendon, a realtime MJO monitoring index and associated computing method were designed, and then, an MJO monitoring and prediction operation system was built up in National Climate Center (NCC). The monitoring results from NCC are much consistent with that from the foreign operational departments (e.g. Australia Bureau of Meteorology). Monitoring results can give a relative good description of the MJO’s intensity and propagation. The realtime MJO prediction operation is also set up using two statistical forecast models. Analysis on the prediction skill showed that two statistical methods give a skillful forecast within 15 days. Lag linear regression model (PCL) has better forecast skill than autoregressive model (ARM).