Megakaryocytes (MKs) undergo a fascinating endomitosis process to become naturally large multilobulated, polyploid nuclei cells that give rise to platelets. Two key steps involving in this process are the controlled failure in cytokinesis (cytoplasm division) and karyokinesis (nuclear division). The cytokinesis defect has been reported to be regulated by factors related to abnormal actomyosin ring and cleavage furrow formation such as RhoA, MYH9 and Rac1. In contrast, it remains largely unknown how nuclear division defect is regulated during MK endomitosis. Using a mouse model and an in vitro differentiation system, we found SETD2 expression markedly declined when MKs underwent polyploidization. Knockdown of SETD2 increased the high ploidy of primary bone marrow (BM) murine MKs, as well as MKs derived from cultured human cord-blood CD34+ cells and MEG-01 cells. Furthermore, the higher ploidy of MKs was accompanied by the increase of platelet formation both in the mouse model and the in vitro differentiation system. With forced overexpression of SETD2, the proportion of high ploidy MKs (≥8N) was remarkably decreased. These data suggest that SETD2, an α-tubulin methyltransferase, negatively regulates polyploidization during megakaryopoiesis. To investigate how SETD2 regulates endomitosis, we employed live-cell imaging to monitor the whole process of endomitosis at the single-cell level by observing the spindle geometry. We found SETD2 deficient MKs were capable of completing the endomitotic cycle, allowing cells to reach a higher ploidy in a significantly shorter time compared to control cells (2.9 hours vs. 4.9 hours). In comparison, overexpression of SETD2 significantly increased endomitosis cycle time compared to control conditions (5.2 hours vs. 4.3 hours). These results show that SETD2 deficiency accelerates the endomitotic cell cycle. Moreover, loss of SETD2 increased the percentage of cells within the presence of nucleoplasmic bridges (NPBs), a thin DNA link referred as a characteristic structure for karyokinesis defect. These results suggest that SETD2 may regulate endomitosis through facilitating the defect of karyokinesis. During karyokinesis, microtubules localized near spindle poles are need to be depolymerized as force-generating apparatus to separate nuclei. Thus, we measured the amount of polymerized and depolymerized microtubule forms in the presence and absence of SETD2. We found that loss of SETD2 decreased the proportion of the depolymerized form of tubulin in spindles, while SETD2 overexpression increased depolymerized microtubule forms. These data suggest that loss of SETD2 promotes the defect of karyokinesis by reducing depolymerized microtubule state. Mechanistically, as SETD2 could trimethylate α-tubulin at lysine 40 (α-tubK40me3), we found that the levels of α-TubK40me3 near spindle poles were inversely correlated with the degree of karyokinesis defect and MK polyploidization. Moreover, SETD2-deficiency resulted in an absence of α-TubK40me3 with an increase of the MK ploidy. And this ploidy increase was largely reversed by overexpressing a mutant tubulin protein that mimics the α-tubK40me3. Together, these data indicate that SETD2-mediated α-TubK40me3 regulates MK polyploidization by influencing microtubule dynamics. In summary, our study reveals that SETD2 is an important, novel regulator of MK endomitosis by regulating α-TubK40me3 to influence microtubule dynamics. These findings help fill current knowledge gaps in the mechanism of the karyokinesis defect during MK endomitosis. Targeting SETD2 may improve high ploidy MKs and platelet production that could promote the regeneration of platelets in vitro and help treat megakaryocytic dysplasia diseases with abnormal polyploidization in the future.