The dissertation main subject was focused on neural modeling in the pharmaceutical technology and biopharmaceutics. Neural models of drugs dissolution profiles and microemulsions compositions were created in order to assess abilities of neural networks to generalize new pharmaceutical formulations. Necessary computer programs were also projected and written using Borland Delphi 5.0 language. It was found that neural networks sets (experts committees) were superior to single neural networks. Special systems ME_expert and SD_expert were constructed in order to carry out in silico screening for microemulsions and solid dispersions respectively. Both systems allowed for generalization of new formulations. Prediction performance of new compositions of microemulsions was estimated on 77%. PKB 2.0 system was also constructed. It allows for approximation of whatever-kind of time-concentration profile for in vitro and in vivo assays. It was found that neural approximation, although less precise, may be equal to regression method. PKB 2.0 was successfully applied in data preprocessing for neural models of drugs dissolution from solid dispersions. Neural preprocessing was beneficial for interpolation ability of abovementioned models. The dissertation proved universal character of artificial neural networks in the context of their application as modeling tools in pharmaceutical technology and biopharmaceutics.