This study has been conducted to determine the effects of (on job) IT security training on the management of computer network's security within Kenyan Public Universities. Computer networks in Kenyan institutions have in the recent past experienced a number of security compromises, leading into huge financial losses and breaches of system integrity. Hacking communication channels is a daily occurrence in Kenya. Findings by Deloitte Kenya Ltd indicated that East African business computer networks are still vulnerable to attack, fraud and confidentiality breaches. To reduce the problem of hacking computer networks, institutions conduct on job- Information Technology security training for IT professionals. While some people perceive the training to bring forth positive returns, others feel it may expose the internal IT systems to even more risks. Despite the uncertainty, institutional managers continually invest heavily on IT security training. This has led to a focused attention in on job IT security Training and its effects on computer network security within the Kenyan Public Universities, to analyze how its management is affected by this training. The specific objectives have been; to determine the effects of on job IT security training on the management of network access control / monitoring, to determine the effects of on job IT security training on the management of network data security, to determine the effects of on job IT security training on the management of network malware control and to determine the effects of on job IT security training on the use of network security policy. The research has been conducted using a quasi-design model of non¬equivalent groups to determine performance of the four network security elements in Kenyan public Universities. Out of the 31 public Universities in Kenya, with the population of 409 IT techies, a sample of 21 Universities has been used to give 150 IT staff members, chosen randomly for administering the questionnaire to collect data. From each sampled University, 37percent of the population was randomly chosen giving a total of 149 respondents. The primary data was presented in tables and charts and summarized using the arithmetic Mean. Data has been analyzed using correlation and regression model in Tobin's Q equation in conjunction with Likert model. The major outcome of the study has been a positive correlation between the training and all the elements of computer network security management, thus determining the relationship under study. The findings could be significant to organizational policy makers, security trainers and IT heads in managing University network security more effectively.