Disclosed is a method for implementing precomputation of a large number in an embedded system. A modulo module, a modulo adding module, and a Montgomery modular multiplier are invoked according to a data format of a modulus length and a value of each data bit of a binary number corresponding to the modulus length, to perform an iterative operation, so that a precomputation result of a large number can be obtained when the modulus length is an arbitrary value, thereby improving the data processing speed.