Platelets play roles in malignancy, wound healing, and immunity. Nevertheless, their significance in postoperative outcomes is not established. This is a retrospective cohort study of 100,795 patients undergoing cancer surgery in 2010 and 2014 in >500 hospitals. Patients were stratified into five groups based on preoperative platelet counts. Multivariable logistic regression was used to determine the risk of 30-day mortality, morbidities, readmission, and prolonged hospitalization using the mid-normal group as a reference. We adjusted for demographic variables, comorbidities, and operation complexity. In the 2014 cohort, multivariable analysis showed that mortality was higher in patients with thrombocytopenia (OR 1.49, 95% CI [1.23-1.81]), high-normal platelets (OR 1.29, [1.06-1.55]), and thrombocytosis (OR 1.78, [1.45-2.19]). Composite postoperative morbidity followed a similar trend with thrombocytopenia (OR 1.34, [1.25-1.43]), high-normal counts (OR 1.41, [1.33-1.49]), and thrombocytosis (OR 2.20, [2.05-2.36]). Concordantly, the risks of prolonged hospitalization and 30-day readmission followed the same pattern. These results were validated in a large colon cancer cohort from the 2010 database. In conclusion, platelet count is a prognostic indicator in cancer surgeries. This could be related to the role of platelets in wound healing and immunity on one hand, and propagating malignancy on the other.