We investigated the effects of ovariectomy (OVX) and 17 beta-estradiol (0.18 mg per pellet) treatment on spatial learning and memory, hippocampal beta amyloid (A beta) levels, and amyloid plaque counts in double transgenic mice (A/P) carrying mutated amyloid precursor protein (APPswe) and presenilin-1 (PS1-A246E). After OVX at 3 months of age, the mice received estrogen treatment for the last 3 months of their lifetime before they were killed at 6, 9, or 12 months of age. Estrogen treatment in A/P OVX mice increased the number of correct choices in a position discrimination task in the T-maze, and slightly improved their performance in a win-stay task (1/8 arms baited) in the radial arm maze (RAM). However, estrogen treatment did not reverse the A beta-dependent cognitive deficits of A/P mice in the water maze (WM) spatial navigation task. Furthermore, ovariectomy or estrogen treatment in OVX and sham-operated A/P mice had no effect on hippocampal amyloid accumulation. These results show that the estrogen treatment in a transgenic mouse model of Alzheimer's disease (AD) improves performance in the same learning and memory tasks as in the normal C57BL/6J mice. However, the estrogen effects in these mice appeared to be unrelated to A beta-induced cognitive deficits. Our results do not support the idea that estrogen treatment decreases the risk or alleviates the symptoms of Alzheimer's disease by inhibiting the accumulation of A beta or formation of amyloid plaques.