The invention relates to the field of control techniques of copper smelting processes, and in particular to a soft-measuring method for the melt temperature in an oxygen bottom-blown copper smelting process. The method disclosed by the invention comprises the steps of firstly selecting auxiliary variables, carrying out data preprocessing on each auxiliary variable, building a soft-measuring model based on a Boosting PLS (Partial Least Squares) method and correcting the soft-measuring model by adopting a sliding-window method. The method disclosed by the invention has the advantages that the melt temperature which is difficult to detect on line and very important can be measured, so that a necessary condition is provided for the advanced controlled implementation of the oxygen bottom-blown copper smelting process; and on-line correction can be carried out on an existed model by utilizing the soft-measuring model correction to ensure that the model has a self-adaptive capacity and is better adapted to the change of operating conditions.