Experimental determination of aqueous solubility is affected by long time invested and consumption of considerable amounts of sample. To solve these problems, QSPR studies have been applied, in order to establish quantitative relationships between the structure and some property of the molecule of interest, by means of a function capable of predicting a certain quality of a compound. Considering that solubility of orally-administered drugs is influenced by the strongly acid pH of the gastrointestinal tract, obtaining predictive models based on pH is of great interest, which was the purpose of the present study. Then, some computer programs were used: ACDLabs (construction of the training series), MODESLAB (calculation of molecular descriptors), IBM SPSS Statistics (data reduction) and BuildQSAR (obtaining and optimization of predictive models). Finally, 24 mathematical models (M) for the prediction of the aqueous solubility of organic compounds of pharmaceutical interest were defined, by ranges of pH (1-1.3, 1.4-1.7, 1.8-2.1 and 2.2-2.5) and by group of chemical behavior. The relative simplicity along with correlation coefficients and standard errors of estimate close to the unity, suggest the external evaluation of models M3-M5 and M11-M22 and their subsequent use as part of the design and development of orally-administered medications.