One of the major thrusts in organization studies has been directed toward understanding and predicting the employee job performance, which consists of various task-related activities. In recent years there has been increasing interest in exploring performance-related behaviors that go beyond the assigned tasks and responsibilities for which employees are typically held accountable. These behaviors are often referred to as non-task behaviors. Two independent streams of research have been developing. On the positive side, researchers have concerned voluntary behavior that goes beyond the core tasks, which is commonly called organizational citizenship behavior(OCB) or contextual performance. On the negative side, there is an independent stream of researches concerning voluntary, potentially destructive or detrimental acts that may even hurt colleagues or organizations, which are studied under the titles of counterproductive work behavior(CWB) or workplace deviance. Recently, many researchers have concerned the main and mediating effect of emotion on these non-task behaviors. According to these researches, there are many events and situations in the work environment that induce and provoke emotional reactions of employee, either positive or negative. Both forms of emotions produce action tendencies and intentions to act. In general, negative emotions tend to produce CWB and positive emotions tend to produce OCB. Many researches on emotional labor have provided evidence for service workers holding a heightened risk of experiencing negative emotions at work. Frequently service workers may find themselves in a state of emotional dissonance, characterized by behaviors that do not concur with their actual affective states. We can expect that emotional dissonance, as a kind of negative emotions that are usually experienced in the service work to meet clients’demands, can also influence negatively to the non-task behaviors of service employee. Thatis, it can reduce their altruistic or helpful behavior such as OCB and increase to the intention to act many deviant behaviors such as CWB. In addition, many studies have documented a lot of cases in which emotional intelligence have been found to have a positive effect on the job related outcomes, such as job satisfaction and job performance. Emotional intelligence is usually conceptualized as a set of abilities that includes the abilities to perceive and regulate emotions in the self and others, and to understand and use emotions to facilitate job performance. Some other researchers assert that emotional intelligence, as a psychological resource, can also buffer and alleviate the negative effects of many stressful job contexts. In this respect, we can also expect that emotional intelligence will have a main and moderating effect in the relationships between emotional dissonance and non-task behaviors of service employee. The purpose of this study is to test these theoretical expectations empirically. Specifically,this paper is aimed to examine the impacts of emotional dissonance and emotional intelligence of service employee on their non-task performance. In this study, two kinds of non-task behavior are considered. One is an altruistic behavior which has been known as the most important dimension of OCB, and the other is CWB which has been usually classified into two subcategories, interpersonal and organizational deviance. In addition, this paper tests a buffering hypothesis that emotional intelligence will moderate the negative impact of emotional dissonance on the non-task behaviors. Data were collected from 264 service employees of 15companies and a series of hierarchical regression analyses were used for the test of hypotheses. Findings are as follows. As expected in our main hypotheses, perceived emotional dissonance of service employee significantly increased their CWB, whereas it was not significantly related to their altruistic behavior. By contraries, emotional intelligence of service employee did not significantly decrease their CWB, but it explained a significant amount of variance in predicting their altruistic behavior. In addition, the buffering hypothesis of emotional intelligence was partially supported. We found that emotional intelligence of service employee played an important role in attenuating the negative impact of emotional dissonance on their altruistic behavior. But we found no such moderating effect of emotional intelligence on their CWB. The implications of these findings for the management of organization and the limitations of this study are then discussed.