Polyglutamine expansion in certain proteins causes neurodegeneration in inherited disorders such as Huntington disease and X-linked spinobulbar muscular atrophy. Polyglutamine tracts promote protein aggregation in vitro and in vivo with a strict length-dependence that strongly implicates alternative protein folding and/or aggregation as a proximal cause of cellular toxicity and neurodegeneration. We used an intracellular polyglutamine protein aggregation assay based on fluorescence resonance energy transfer (FRET) to identify inhibitors of androgen receptor (AR) aggregation in three libraries of biologically active small molecules: the Annotated Compound Library, the NINDS Custom Collection and a kinase inhibitor collection. In the primary screen 10 compounds reduced AR aggregation. While 10/10 also reduced huntingtin (Htt) exon 1 aggregation, only 2/10 reduced aggregation of pure polyglutamine peptides. In a PC-12 model 9/10 compounds reduced aggregation. Five out of nine compounds tested in an Htt exon 1 assay of neurodegeneration in Drosophila partially rescued the phenotype. Three of the five compounds effective in flies are FDA-approved drugs. These compounds provide new leads for therapeutic development for the polyglutamine diseases based on their efficacy in mammalian cells and a Drosophila model. The high predictive value of the primary screen suggests that the FRET-based screening assay may be useful for further primary and secondary screens for genes or small molecules that inhibit polyglutamine protein aggregation.