Data observed by telemetric system often carries outliers resulting from instrument malfunctioning, false signal acquisition because of signal leak, collision and disturbance during signal transmission, and special measuring demand of hydrologic variable, in addition to unavoidable random errors. When the parameters of real-time co-correction model are estimated by the abnormal data, the algorithm must not only be able to resist the effect of the outliers, but also have ability for real-time tracing of the changes of parameters. In this paper, a robust recursive least-squares algorithm with a forgetting factor is produced based on the recursive least-squares algorithm. And a example is given to demonstrate that the algorithm is insensitive to the outliers and is adapt to the time-varying parameter estimation.