In the present study we validated a widely used, high-throughput in vitro permeability model (PAMPA) to be used at the early stage of drug discovery for the phospholipidosis (PLD) prediction of drug-like compounds. Regarding the mechanism of action of PLD, our pH-gradient PAMPA system is the first noncell based model to mimic one-way transport of cationic amphiphilic drugs (CADs) from cytosol to the lysosome. Moreover, due to the fact that PLD can mainly occur in lung, liver, brain, kidney and heart, we have used similar commercially available original tissue-derived lipid fractions (heart, liver, brain), and in the case mimicking membrane of kidney and lung tissue we prepared tissue-mimetic artificial lipid mixtures in house. Metabolism of a drug can change the degree of PLD depending on the physicochemical properties of metabolites and the rate of metabolism. Our data from 57 drugs and 4 metabolites of earlier and 2 metabolites of newly recognized outliers (phenacetin and bupropion) using our pH-gradient PAMPA system show a good correlation with in vivo PLD data. Moreover, predictive ability of our best system, the lung specific pH-gradient PAMPA model was significantly better than widely used in silico models and it was also slightly better than that of the known noncell based models on our selection of compounds. Our pH-gradient PAMPA systems therefore offer mechanistically alternative, accurate and cost-effective screening tools for the early prediction of PLD potential of drug-like compounds.