Abstract: This study proposes an adaptive information discovery framework for computational grid, called PIVOT. With an active information discovery mechanism, PIVOT can extract and provide explicit information of distributed grid resources for specific scheduling algorithm. By introducing a tunable α-hops flooding method for distributed information query and collection, PIVOT supports customized resources information retrieval to fulfill requirements of applications. The scalable and adaptive information discovering mechanism of PIVOT is better than traditional pre-configured information services. PIVOT is implemented in the grid environment MASSIVE and is evaluated with an actual scheduling algorithm. Experiments demonstrate that PIVOT improves the effectiveness of resources scheduling and lessen the executing time of grid tasks.