This paper presents an agent-based model of labor market to investigate the relationship between company and worker. Unlike most of previous studies of labor market we apply a game theoretic approach to defining entities in labor market: companies and workers. A company can choose the level of wages, and workers can select the level of effort to increase the productivity in response to the wages. Company and worker agents are designed to possess the basic attributes in order to reflect the real labor market and their activities are adaptively changed using evolutionary model. Our approach is illustrated with four simulation results: the effect of workers resignation, sick leave, dismissal of companies and productivity growth. Various experiments were conducted, and the interactions between worker and company are analyzed. Especially performance-based reward strategy and non-greedy strategy in job changing are necessary for companies and workers according to the simulation results. The experimental result confirms that the balanced power between worker and company is important in maintenance and extension of labor market. Nash Equilibrium can be maintained in all cases.