A pre-arranged power failure time predication method and system are provided. The method comprises the following steps of: acquiring screened production control indicators used for predicating, and acquiring a historical data sample; establishing a linear relation model of the production control indicators and pre-arranged power failure time according to the production control indicators; determining a regression coefficient according to the historical data sample; substituting the regression coefficient into the linear relation model, and determining a predicated value of the pre-arranged power failure time; and the production control indicators comprising a load supply capability type indicator, a network structure level type indicator, an operation management level type indicator, an equipment technical level type indicator and a capital construction technical modification investment type indicator. The method and system consider the plurality of production control factors including a supply capability, a network structure level, an operation management level, an equipment technical level and capital construction technical modification investment and the like, and the predicated value of the determined pre-arranged power failure time is more reliable and reasonable.