Tumor growth in mice preclinical prostate cancer model (human prostate cancer cells, PC-3) was followed for 4 weeks by weekly DWI in control group (n=10) and treatment group (n=9) receiving Docetaxel. DWI data sets were acquired using 12 b-values the range of 0-2000 s/mm2. The DWI signal decays were fitted using monoexponential, biexponential, kurtosis and stretched exponential models/functions. Independent least squares fitting and spatially constrained Maximum Penalized Likelihood Estimation have been applied. The spatially constrained Maximum Penalized Likelihood Estimation revealed the effect of treatment in mice subjects, while conventional LSQ fitting failed to reveal significant difference between control and treatment.