For hardware regular expression matching algorithms are faced with the challenges in memory space and throughput, a Predefined-class Compact Finite Automaton (Pre-class CFA) matching algorithm based on Extended Finite Automaton (XFA) regular expression matching algorithm was proposed. By using the predefined classes, the algorithm not only can achieve class character matching in regular expression, but also can match the special character set by priority setting; besides, the migrating edges were reduced when eliminating the Deterministic Finite Automaton (DFA) state space exploration problem. At the same time, based on the characteristics of Field-Programmable Gate Array (FPGA) and migration edge, a migration edge access method based on parallel Read-Only Memory (ROM) was designed to realize the parallel reading and matching of the same state with mutiple migration edges. The algorithm was tested on ALTERA DE2-70 FPGA platform, which is a low-performance platform. The throughput of the experimental system is 1.30 Gb/s, which can realize intrusion detection and garbage filtering under gigabit network.