The recursive least-squares algorithm was adopted to investigate the estimation of surface heat flux of inverse heat conduction problem from experimental data. The Kalman filtering technique which accounted for the residual innovation sequence and the least-squares estimation which accounted for computing heat flux was introduced to treat one-dimensional inverse heat conduction problem. By virtue of recursive algorithm, an on-line estimation can be made in place of batch form off-line estimation. The method is adequate for impulse heat flux estimation.