The objective of this study was to prepare and evaluate rate-controlled 17beta-estradiol nanoparticles (E2-NPs), and then optimize them by central composite design (CCD) using a response surface methodology (RSM). The E2-NPs were designed for cerebral ischemia therapy and prepared by oil-in-water (o/w) emulsion by mixing E2 and phosphatidylcholine, cross-linking of the bovine serum albumin (BSA) wall material with glutaraldehyde, and subsequently creating a biodegradable coating shell with L-arginine. In two-factor, five-level CCD, the independent variables were the pH value of the aqueous phase (X1) and the amount of glutaraldehyde (X2). The optimal formulation of the E2-NPs was developed from the response equations of each fitted model. The optimal E2-NPs particle size was 92.4 +/- 1.4 nm, and its cumulative release in 4.0 h (therapeutic window for stroke) was 71.17 +/- 0.24%, with a zero-order release model of 24 h. The preparation and testing of the optimal formulation showed a good correlation between the predicted and observed values. In addition, a brain microdialysis study demonstrated that the E2-NPs had the ability to penetrate the blood-brain barrier and clearly increase and maintain the drug concentration in the brain over 12 h. Furthermore, the E2-NPs reduced brain infarct size in rats with middle cerebral artery occlusion (MCAO)-induced stroke. These results suggest that rate-controlled E2-NPs increase the efficacy of E2 for ischemic stroke therapy.