In a complex industrial batch processes such as the top-blown BOF steelmaking process, it is a complicated task to monitor and act on the progress of several important control parameters in order to avoid an undesired process event such as "slopping" and to secure a successful batch completion such as a sufficiently low steel phosphorous content. It would, therefore, be of much help to have an automated tool, which simultaneously can interpret a large number of process variables, with the function to warn of any imminent deviation from the normal batch evolution and to predict the batch end result. One way to compute, interpret, and visualize this "batch evolution" is to apply multivariate data analysis (MVDA). At SSAB Europe's steel plant in Luleå, new BOF process control devices are installed with the purpose to investigate the possibility for developing a dynamic system for slopping prediction. A main feature of this system is steelmaking vessel vibration measurements and audiometry to estimate foam height. This paper describes and discusses the usefulness of the MVDA approach for static and dynamic slopping prediction, as well as for end-of-blow phosphorous content prediction. Multivariate data analysis (MVDA) methods have been applied on the top-blown BOF steelmaking process, with the main aim to create industrially applicable static (i.e., prior to blow), as well as dynamic in-blow batch models for predicting the slopping probability. The MVDA approach has also been investigated in regard to in-blow prediction of end-of-blow phosphorous content.