Abstract: Burden distribution in a blast furnace was estimated based on least squares approximations and multi-radar data. Firstly, a three-segment curve, which includes two straight lines and a quadratic curve, was used to describe the burden distribution. Secondly, based on the burden distribution principles, some constraint equations were obtained to estimate parameters in the three-segment curve, which makes the burden surface profile more reasonable. Then, the burden distribution was estimated by using least squares approximations and multi-radar data, and a real-timely display of the burden surface profile could be achieved. A numerical example with real multi-radar data obtained from a steel plant shows the effectiveness of the estimate method.