Influenza virus A and B are common pathogens that cause respiratory disease in humans. Recently, a highly virulent H5N1 subtype avian influenza virus caused disease outbreaks in poultry around the world. Drug resistant type A viruses rapidly emerged, and the recent H5N1 viruses were reported to be resistant to all current antiviral drugs. There is an urgent need for the development of new antiviral drugs target against both influenza A and B viruses. This dissertation describes work to identify small molecule inhibitors of influenza protein NS1 by a high throughput fluorescence polarization assay. The N-terminal GST fusion of NS1A (residue 1-215) and NS1B (residue 1-145) were chosen to be the NS1A and NS1B targets respectively for HT screening. In developing the assay, the concentrations of fluorophore and protein, and chemical additives were optimized. A total of 17,969 single chemicals from four compound libraries were screened using the optimized assay. Six true hits with dose-response activity were identified. Four of them show an IC₅₀ less than 1 [micromolar]. In addition, one compound, EGCG, has proven to reduce influenza virus replication in a cell based assay, presumably by interacting with the RNA binding domain of NS1. High throughput, computer based, virtual screenings were also performed using four docking programs. In terms of enrichment rate, ICM was the best program for virtual screening inhibitors against NS1-RBD. The compound ZINC0096886 was identified as an inhibitor showing an IC₅₀ around 19 [micromolars] against NS1A, and 13.8 [micromolars] against NS1B. In addition, the crystallographic structures of the NS1A effector domain (wild type, W187A, and W187Y mutants) of influenza A/Udorn/72 virus are presented. A hypothetical model of the intact NS1 dimer is also presented. Unlike the wild type dimer, the W187Y mutant behaved as a monomer in solution, but still was able to binding its target protein, CPSF30, with wild type binding affinity. This mutant may be a better target for the development of new antiviral drugs, as the CPSF30 binding pocket is more accessible to potential inhibitors. The structural information of those proteins would be very helpful for virtual screening and rational lead optimization. / text