The exponential increase of malicious and criminal activities in cyber space is posing serious threat which could destabilize the foundation of modem information society. In particular, unexpected network paralysis or break-down created by the spread of malicious traffic could cause confusion and disorder in a nationwide scale, and unless effective countermeasures against such unexpected attacks are formulated in time, this could develop into a catastrophic condition. As a result, there has been vigorous effort and search to develop a functional state-level cyber-threat early-warning system however, the efforts have not yielded satisfying results or created plausible alternatives to date, due to the insufficiency of the existing system and technical difficulties. The existing cyber-threat forecasting and early-warning depend on the individual experience and ability of security manager whose decision is based on the limited security data collected from ESM (Enterprise Security Management) and TMS (Threat Management System). Consequently, this could result in a disastrous warning failure against a variety of unknown and unpredictable attacks. It is, therefore, the aim of this research to offer a conceptual design for "Knowledge-based Real-Time Cyber-Threat Early-Warning System" in order to counter increasinf threat of malicious and criminal activities in cyber suace, and promote further academic researches into developing a comprehensive real-time cyber-threat early-warning system to counter a variety of potential present and future cyber-attacks.