This paper aims at describing the use of two well-known techniques, i.e. expert systems and soft sensors in order to develop a prevention method for the froth collapse problem. This will be achieved by monitoring concentrate froth solid-to-liquids ratio. This scheme is outlined like an alternative solution to artificial vision based methods, and it was tested by simulation under several perturbations affecting the metallurgical performance. (c) 2004 Published by Elsevier Ltd.