During development, metaphase spindles undergo large movement and/or rotation to determine the cell division axis. While it has been shown that spindle translocation is achieved by astral microtubules pulling and/or pushing the cortex, how metaphase spindle stability is maintained during translocation remains not fully understood. In budding yeast, our lab has previously proposed a model for spindle orientation wherein the mitotic spindle protein She1 promotes spindle translocation across the bud neck by polarizing cortical dynein pulling activity on the astral microtubules. Intriguingly, She1 exhibits dominant spindle localization throughout the cell cycle. However, whether She1 has any additional role on the spindle during translocation is unknown. Using function-separating alleles, live-cell fluorescence microscopy, biochemical and biophysical assays, I found that She1 ensures metaphase spindle integrity by crosslinking and stabilizing interpolar microtubules. Loss of this stabilizing activity leads to spindle defects characterized by an increase in interpolar microtubule turnover rate and a pronounced spindle bending/collapse phenotype. She1 binds microtubules via its C-terminal domain, and crosslinks microtubules possibly by self-association. The N-terminal domain contains a tandem newly identified PISH1/2 motif that can regulate Ipl1/Aurora targeting to the mitotic spindle, which, in turn, negatively regulates She1 crosslinking activity and localization on the spindle. My work reveals a new role for She1 in maintaining spindle integrity and offers insight into how metaphase spindles could be stabilized by a non-motor crosslinking microtubule-associated protein (MAP) during dynein-dependent spindle positioning in higher organisms.