Summary:A computer program for phenytoin (PHT) dosing was developed containing seven different menus: two for drug‐naive patients, one using an empirical equation, the other using means forVmaxandKm; two for patients in whom either one or two dose rates and steady‐state concentrations are available; two for patients with hypoalbuminemia, and uremia, respectively; and one menu that optimizesVmaxandKmfrom available steady‐state concentrations. The program accepts or converts PHT and sodium PHT, and makes blood level correction for the concomitant administration of 25 different drugs. The evaluation of the program was done by retrospective analysis using data from three study pools: group I involved 47 patients from the University Hospital, group II relied upon 29 patient data supplied from a collaborative Veterans Administration study, and group III involved 26 patients from the Children's Hospital. Predictions were made and compared with found data to be within a range of ± 15, 20, or 25%. For study group III, many individual blood samples were <8 μg ml−1; hence, saturation kinetics may not have been involved. It is suspected that saturation kinetics in infants may begin at higher levels. Compliance seems to still be a major problem in PHT monitoring and dosage regimen adjustment. Accepting the data as they are, using one or two dose rates with the corresponding blood concentrations resulted overall in 73‐86% achieving blood levels within ± 25% of the predicted value.