The level of trust in a nation is treated as the most pervasive cultural characteristic influencing the nation’s prosperity and competitiveness. Also in the field of organization studies, the importance of trust has been argued to reduce transaction costs within organizations, increase spontaneous sociability among organizational members, and facilitate appropriate forms of deference to organizational authorities. Therefore, trust within an organization is a socially complex resource which contributes to the sustaining of competitive advantages. However, organizational trust-building is not an easy work due to the conceptual characteristic of trust and the multi-dimensions of organizational relationship. Trust is basically a psychological state of a trustor which is being developed in the interdependent relationship with a specific trustee. There may be various relationships in organizations (i.e., person-person, person-group, person-organization, group-group, group-organization, and/or organization-organization) which need trust-building. Any of substantial entities (i.e., person, group, organization) composing an organization can be a trustor and/or a trustee. So, high level of trust in a specific relationship does not mean high level of trust within the organization. Therefore, trust-building as a managerial issue needs a specific dyad to be identified. This study focuses on the two-way trust (i.e., leader’s trust in subordinate and subordinate’s trust in leader) in leader-subordinate dyad as determinants of subordinate’s trust in organization. Trust in leader-subordinate relationship has been studied by many researchers, but trust in organization-member relationship has been paid little attention to. The established models of interpersonal trust fundamentally assume that one party's trust in another party reciprocally reinforce the latter’s trust in the former, but research efforts to support this reinforcement process of reciprocal trust are not sufficient. Further, despite much research in subordinate’s trust in leader, almost no attention has been paid to leader's trust in subordinate. In this context, major research questions of the study are whether a subordinate trust his/her leader more as he/she perceives trust from his/her leader, whether a subordinate trust his/her organization more as he/she trust his/her leader or as he/she perceives trust from his/her leader based on the perspective that leaders are representatives of an organization to subordinates, and finally whether subordinate’s trust in leader mediates the relationship between leader’s trust in subordinate and subordinate’s trust in organization. To collect the data, we conducted questionnaire-based survey for all pilots of Korean Air on May, 2006. 1,600 questionnaires were sent to the individual mail box of the pilots and 233 of them responded (response rate =14.5%). However, the data of 182 respondents were used for analysis because 51 with missing values were excluded. The model of this study includes an independent variable, a mediator, a dependent variable, and three control variables. They are leader’s trust in subordinate(LTS), subordinate’s trust in leader(STL), subordinate’s trust in organization(STO), and propensity to trust(PT) and two demographic variables, respectively. All variables except for demographics were measured by a Likert-type 7-point scale where 1, 4, and 7 meant “strongly disagree,” “medium,” and “strongly agree,” respectively. The Cronbach’s alpha values of LTS, STL, STO, and PT were 0.96, 0.95, 0.87, and 0.84, respectively. We conducted an exploratory factor analysis for a validity test. Then we computed the average score of all the items of each variable and used them for correlation analysis and multiple regression analyses. To test the mediation effect hypothesized, we used the hierarchical multiple regression analyses which Baron and Kenny recommended and checked the result of a Sobel test. We also diagnosed the possibility of common method bias in the sample with Harman’s single factor test and multicollinearity problem with variance inflation factor and condition index. The findings have shown that the more trust subordinates perceive from their leader, themore they trust their leader (LTS-STL); that the more trust subordinates perceive from their leader the more they trust their organization (LTS-STO); that the more subordinates trust their leader the more they trust their organization (STL-STO); and that subordinates’ trust in leader mediates the relationship between LTS and STO. From the findings, we would propose three major implications or contributions. First, this study contributes to a theoretical development by suggesting one possible explanation to link LTS and STL in leader-subordinate dyad. Second, our research gives an answer to the question of why leaders should trust subordinates first rather than waiting for their trust, based on the premise that risk-taking(or trusting) behavior should be distinguished from trustworthiness. Lastly, our research suggests the new role of middle manger in building trust within an organization. From the finding that subordinate’s perception of LTS is positively related to STO and mediated by STL, it becomes clear that the top managers of organizations should put priority on developing organizational norms and practices motivating middle managers to proactively trust their subordinates as well as to increase his/her own trustworthiness.