In Grid, cooperation often occurs between unknown entities. To guarantee smooth and reliable ongoing of such cooperation, reliable trust relationship must be established between them. With wide application in Electronic Commerce and online communities, Reputation mechanism emerges as a promising solution, where a scientific reputation evaluation method is crucial. Yet, most methods currently available do not consider Grid's distinct characteristics such as the sparseness of ratings and strangeness of entities, which are not feasible to Grid. In this paper, we propose a bias-tuned dishonestyresistant reputation evaluation method for trust establishment in Grid. What distinguishes our trust model is the introduction of a pre-evaluating set, with which we can reasonably filter out malicious ratings, effectively tune a rater's bias to cater to the current evaluator's criteria and scientifically weight a rater's rating in aggregation.