Methods: (i) Animal experiments. This study conducted experiments using specific pathogen-free (SPF) grade male C57BL/6J wild-type (WT) mice and APP/PS1 double transgenic mice. The animals were divided into three groups: WT group (WT mice, n = 5, receiving distilled water daily), APP/PS1 group (APP/PS1 double transgenic mice, n = 5, receiving distilled water daily), and BSTSD group [APP/PS1 double transgenic mice, n = 5, treated with BSTSD suspension at a dosage of 27 g/(kg·d) for 90 d]. Cognitive function was assessed using the Morris water maze (MWM). Post-experiment, hippocampal tissues were collected for analysis of pyramidal cell and synaptic morphology through hematoxylin-eosin (HE) staining and transmission electron microscopy (TEM). (ii) Cell experiments. The HT-22 cells were divided into control group (untreated), Aβ25-35 group (treated with 20 μmol/L Aβ25-35 for 24 h), icariin group (pre-treated with 20 μmol/L icariin for 60 min, followed by 20 μmol/L Aβ25-35 for an additional 24 h), and icariin + LY294002 group [treated with 20 μmol/L icariin and 20 μmol/L LY294002 (an inhibitor of the phosphoinostitide 3-kinases (PI3K) signaling pathway) for 60 min, then exposed to 20 μmol/L Aβ25-35 for 24 h], and cell viability was measured. Western blot was used to detect the expression levels of synapse-associated proteins [synaptophysin (SYP) and postsynaptic density-95 (PSD-95)] and PI3K signaling pathway associated proteins [phosphorylated (p)-PI3K/PI3K, p-protein kinase B (Akt)/Akt, and p-mechanistic target of rapamycin (mTOR)/mTOR]. Results: (i) Animal experiments. Compared with APP/PS1 group, BSTSD group showed that escape latency was significantly shortened (P < 0.01) and the frequency of crossing the original platform was significantly increased (P < 0.01). Morphological observation showed that pyramidal cells in the hippocampal CA1 region were arranged more regularly, nuclear staining was uniform, and vacuole-like changes were reduced after BSTSD treatment. TEM showed that the length of synaptic active zone in BSTSD treatment group was increased compared with APP/PS1 group (P < 0.01), and the width of synaptic gap was decreased (P < 0.01). (ii) Cell experiments. Icariin had no obvious toxicity to HT-22 cells when the concentration was not more than 20 μmol/L (P > 0.05), and alleviated the cell viability decline induced by Aβ25-35 (P < 0.01). Western blot results showed that compared with Aβ25-35 group, the ratios of p-PI3K/PI3K, p-Akt/Akt and p-mTOR/mTOR in icariin group were significantly increased (P < 0.01), while the protein expression levels of SYP and PSD-95 were increased (P < 0.01). These effects were blocked by LY294002 (P < 0.01). Conclusion BSTSD and icariin enhance cognitive function and synaptic integrity in AD models and provide potential therapeutic strategies through activation of the PI3K/Akt/mTOR pathway.