The feasibility of predicting at laboratory-scale the fouling propensity of orange juice suspended particles based on their physical properties, was investigated. Samples were generated from mixtures of two selected size class of large and fine particles. The particle-size distribution, rheological and textural properties of these samples were characterised as well as their fouling propensity using permeability test. Qualitative and quantitative statistical analyses revealed significant correlations between specific resistance to filtration (SRF) and hydration velocity (Vh), statistical diameter (D90) and shear strain at flow point (¿f). Partial Least Square method allowed to discriminate contrasted values of log(SRF) of suspended particles using a simplified model (R2 =¿0.79; RMSEC; RMSECV <20% of SRF mean; RPD = 1.6), based on 5 predictor variables (Vh, ¿f, D90, D[4;3] and ). These findings, offer new prospects to guide the choice of pretreatment or specific operating conditions to improve the performance of fruit juices microfiltration.