Temporal changes in short-term mortality following surgery for colorectal cancer (CRC) are unknown. We examined temporal changes in 30-day postoperative mortality, as well as changes in preoperative and postoperative disorders that could contribute to 30-day mortality. Using national Veterans Administration (VA) administrative data, we identified patients with CRC during 1987-2000 who received surgical resection. Cox proportional hazards models were used to evaluate the association between the risk of 30-day mortality and year of surgical resection, while adjusting for several preoperative disorders, disease comorbidity, as well as hospital surgical volume. A total of 32,621 patients were identified. The 30-day postoperative mortality declined from 4.7% during 1987-1988 to 3.9% during 1998-2000. Patients who received surgical resection during 1992-1994, 1995-1997, and 1998-2000 had a 14, 14, and 27% lower adjusted risk of 30-day mortality, respectively, compared with those resected in 1987-1988. Preoperative disorders associated with increased mortality included chronic pulmonary disease, congestive heart failure, diabetes, hemiplegia/paraplegia, moderate/severe liver disease, and renal disease. Significant declines were observed in several postoperative disorders including anesthesia complications and thromboembolism. An improvement in 30-day postoperative mortality following surgical resection for CRC was observed. Declining preoperative and postoperative disorders, as well as improvements in surgical care, could partly explain these findings.