Genetic algorithms (GA) were used to develop specific technetium metal-ligand force field parameters for the MM3 force field. These parameters were developed using automated procedures within the program FFGenerAtor from a combination of crystallographic structures and ab initio calculations. These new parameters produced results in good agreement with experiment when tested against a blind validation set. To illustrate the utility of these new force field parameters, quantitative structure-activity relationship (QSAR) models were developed to predict the P-glycoprotein uptake (log10 VI) of a series of hexakis(areneisonitrile)technetium(I) complexes and to predict their biodistribution. The log10 VI QSAR model, built using a training set of 16 Tc(I) isonitrile complexes, exhibited a correlation between the experimental log10 VI and 5 simple descriptors as follows: r2 = 0.94, q2 = 0.93. When applied to an external test set of six Tc(I) isonitrile complexes, the QSAR preformed with great accuracy q2 = 0.78 based on a leave-one-out cross-validation analysis. Further QSAR models were developed to predict the biodistribution of the same set of Tc(I) isonitrile complexes; a QSAR model to predict hepatic uptake exhibited a correlation between the experimental log10(Blood/Liver) with six simple descriptors as follows: r2 = 0.97, q2 = 0.96. A QSAR model to predict renal uptake exhibited a correlation between the experimental log10(Blood/Kidney) and six simple descriptors as follows: r2 = 0.85, q2 = 0.82. When applied to the external test set the QSAR models preformed with great accuracy, q2 = 0.78 and 0.56, respectively.