In Alzheimer's disease (AD), amyloid beta (Aβ) plaques initiates a cascade of pathological events where the overactivation of N-methyl-d-aspartate receptors (NMDA) by excess glutamate (Glu) triggers oxidative processes that lead to the activation of microglial cells, inflammation, and finally neuronal death. Amaryllidaceae alkaloids exert neuroprotective activities against different neurotoxin-induced injuries in vitro, and although their biological potential is well demonstrated, their neuroprotective activity has not been reported in an in vivo model of AD. The aim of our study was to determine the in vitro and in vivo neuroprotective potential of standardized alkaloidal fractions of Zephyranthes carinata. In this work, the neuroprotective effect of two alkaloidal fractions extracted from Z. carinata (bulbs and leaves) was analyzed in an in vitro excitotoxicity model in order to select the most promising one for subsequent evaluation in a triple transgenic mouse model of AD (3xTg-AD). We found that Z. carinata bulbs protect neurons against a Glu-mediated toxic stimulus in vitro, as evidenced by the decrease in apoptotic nuclei, the reduction in the lipid peroxidation product malondialdehyde and the conservation of dendritic structures. The effects of intraperitoneal administration of Z. carinata bulbs (10 mg/kg) every 12 h for 1 month on 3xTg-AD (18 months old) showed improved learning and spatial memory. Histopathologically, the alkaloidal fraction-treated 3xTg-AD mice exhibited a significant reduction in tauopathy and astrogliosis, as well as a significant decrease in the proinflammatory marker COX-2 and an increase in pAkt. The results suggest that Z. carinata bulbs provide neuroprotective effects both in vitro and in 3xTg-AD mice by intervening in the inflammatory processes, regulating the aggregation of pair helical filaments (PHFs) and activating survival pathways.