Due to the complicated causes

the poor understanding to debris flow and difficulties in prediction

a Fisher discrimination analysis model for forecasting the debris flow was developed according to the theory of discrimination analysis.Its basic idea is that a series of factor data related to the debris flow occurrence was projected to some direction and the data projected can be grouped as far as possible.The discriminant function was determined according to the univariate variance analysis

which requires smaller intraclass variance and larger interclass variance.And then new samples were sorted into some class according to this criterion.The proposed model is simple and run fast

and can predict the condition of debris flow in time.The case study showed that Fisher discrimination analysis model can improve the prediction accuracy of debris flow.