Abstract: The soft-sensing model for mill load(ML) is difficult to adapt to the time-varying characters of the mineral process,and it needs to be updated online in real-time according to the changes of condition.Aiming at these problems, based on the vibration spectrum of the mill shell,an on-line soft-sensing approach is proposed to measure the ML parameters, such as material to ball volume ratio(MBVR),pulp density(PD) and charge volume ratio(CVR) inside the mill.The method is realized by the integration of recursive principal component analysis(RPCA) and on-line least square support vector regression(LSSVR).At first,for the training samples,spectral principal components(PCs) at low,medium and high frequency bands of the shell vibration spectrum are extracted through PCA.Then,the spectral PCs of serial combination with different bands are used to construct ML parameters off-line soft sensing models based on LSSVR.At last,when a new sample is given,after predicted with the older models,the inputs and regression parameters of the soft sensing models are updated by RPCA and on-line LSSVR algorithm respectively.Therefore,the on-line updating of the soft-sensing models for ML parameters are implemented.Experiment result shows that the proposed approach has higher accuracy and better predictive performance than other normal approaches.