While regulation of spindle positioning and orientation is known to be mediated by cortical pulling forces generated by dynein on astral microtubules, the contribution of microtubule-based pushing forces remains less clear. We developed a two-dimensional computational model that incorporates both cortical pulling and microtubule pushing to examine spindle behavior across diverse cell types and geometries. The model successfully recapitulates experimental observations in cells that round during mitosis, including Drosophila follicular epithelial cells, Drosophila neuroblasts, and the early C. elegans embryo. Our analysis shows that pulling forces are the dominant determinants of directional spindle orientation. However, pushing forces play essential supporting roles by preventing spindle stalling, enhancing robustness, and facilitating efficient alignment. We further applied our model to irregularly shaped zebrafish endothelial cells, which exhibit complex, non-spherical geometries and dynamic morphology. In this context, our model demonstrates that asymmetric distributions of cortical force generators, potentially enriched at cell-cell junctions, are sufficient to explain off-center spindle positioning. Together, this work establishes a unified framework linking cell geometry, molecular polarity cues, and force balance to spindle dynamics.