Background: The prognosis of endocrine therapy for prostate cancer (PCa) may be related to the micro perfusion characteristics of tumors. The purpose of this study was to compare the prediction and evaluation values of mono-exponential, bi-exponential, and stretched exponential diffusion weighted imaging (DWI) in the efficacy of endocrine therapy for PCa. Materials and Methods: A total of 102 PCa patients who underwent endocrine therapy for 6 months were included in this study. The patients were divided into effective and ineffective groups. Multipara metric prostate MRI including a multi-b (11 b factors) DWI was performed before treatment to measure the standard ADC (ADC_stand) from mono-exponential, slow ADC (D), fast ADC (D*) and perfusion fraction (f) from bi-exponential, distributed diffusion coefficient (DDC) and diffusion heterogeneity index (α) from stretched exponential, respectively. The Mann-Whitney U test, binary logistic regression, restricted cubic spline, and receiver operating characteristic (ROC) curve were used for statistical analysis of the above parameters and prognosis of endocrine therapy for PCa in two groups. Results: Sixty-one cases in the effective group and 41 cases in ineffective group were included in study. Only D* and f had a significant difference between the two groups before treatment (P<0.05). D* (OR 2.417, 95%CI 1.410 to 4.144) and f (OR 0.982, 95%CI 0.972 to 0.992) had a significant influence on the prognosis of treatment (P<0.05). D* was associated with the prognosis of endocrine therapy for PCa in a non-linear manner (P<0.05). According to the ROC curve, D* and f showed high diagnostic significance with the AUC values of 0.932 and 0.885, respectively. Conclusion: D* and f of biexponential before treatment can effectively predict and evaluate the prognosis of endocrine therapy for PCa, which may reflect the micro perfusion characteristics of tumors.