Traditional approaches for facilitating the learning process in risk analysis and management in software projects usually lack practical activities in solving real problems. This is basically due to time restrictions and the difficulty of simulating real scenarios where such risks occur. In this sense, games have been used as a learning resource to simulate real scenarios and allow users to learn and practice with no real risks or losses. In this paper, we present an ongoing project to develop a multi-player, persistent, Web-based game. This game, named eRiskGame, aims to improve the learning process in risk management in software projects based on intelligent agents and fuzzy systems. The intelligent agents control the actions of the game based on fuzzy systems that store expert knowledge in fuzzy rule bases. Users play the role of software project managers who have to compete for market resources by presenting competitive budgets and software management plans. Users also have to deal with real-world problems involving risks, such as managing employees, budgets, and changes in project plans. The game is being validated in two undergraduate software engineering courses. This paper introduces the techniques used in the project and details eRiskGame.