MgNi-based alloys have been regarded as very promising anode material for nickel-metal hydride battery owing to their electrochemcial capacity much higher than that of AB(5) alloys. However, Mg-Ni alloy anodes generally suffer very serious capacity decay in cycling, which is previously ascribed to the corrosion of Mg in alkali electrolyte in the past decades. To further reveal the capacity fading mechanism of MgNi-based alloys, the electrochemcial hydrogen storage properties of Mg2Ni and amorphous Mg50Ni50 alloys and their microstructural evolution during cycling were comparatively studied. It has been firstly demonstrated that it is the amorphous MgNi phase instead of Mg2Ni phase contributes to electrochemical discharge capacity, because the Mg2NiH4 hardly released hydrogen under electrochemcial conditions. Then, it has been found that the charge input level has significant effect on the cyclic performance of amorphous Mg50Ni50 alloy anode. At full charge of 500 mAh g(-1), the amorphous MgNi phase easily crystallized to the nanocrystalline Mg2NiH4, which leaded to the capacity decay immediately. Comparatively, with the low charge input of 100 mAh g(-1), the discharge capacity of amorphous Mg50Ni50 alloy remained almost unchanged for 425 cycles. The excellent cyclability is attributed to the less hydrogenation-induced crystallization at low charge input. This work clearly demonstrates that the hydrogenation-induced crystallization is a dominating reason for the electrochemical capacity decay in the milled amorphous Mg-Ni anode. It provides a new approach to improve the electrochemcial properties of Mg-Ni alloy electrode by inhibiting the hydrogenation-induced crystallization. (C) 2019 Elsevier Ltd. All rights reserved.