There is some uncertainty about pathological grading of mucinous colorectal adenocarcinoma, defined as colorectal cancer demonstrating at least 50% mucinous differentiation. Under the WHO 2000 classification mucinous colorectal cancer was considered high grade. However under the current WHO 2010 classification microsatellite unstable/mismatch repair-deficient (MSI/MMRd) mucinous colorectal cancer is considered low grade, whereas microsatellite stable/mismatch repair proficient (MSS/MMRp) tumours are high grade. However there is little empirical evidence for this approach. We therefore compared the long term survival of patients with MSI/MMRd vs MSS/MMRp mucinous colorectal cancer in a large unselected cohort of patients undergoing surgery at our institution from 1998 to 2011. There were 2608 patients in the cohort, of which 264 (10.1%) were mucinous. 95 (36%) of the mucinous tumours were microsatellite unstable. The all-cause 5-year survival of mucinous MSI/MMRd colorectal cancer was similar to that of non-mucinous low-grade colorectal cancer (73 vs 67%, P = 0.368), and significantly better than that of both non-mucinous high-grade (73 vs 53%, Po 0.001) and mucinous MSS/MMRp colorectal cancer (73 vs 57%, P = 0.023). The 5-year survival of mucinous MSS/MMRp colorectal cancer was slightly better than that of non-mucinous high-grade patients (57 vs 53%, P = 0.027), but significantly worse than that of non-mucinous low-grade colorectal cancer (57 vs 67%, P = 0.018). In multivariate Cox regression analysis, conventional histological grade based on glandular differentiation maintained prognostic significance (P = 0.003) whereas MSI/MMRd status just failed to be statistically significant (P = 0.062). Our findings support the WHO 2010 approach that as a group mucinous MSS/MMRp colorectal cancers are biologically aggressive. However, grading based exclusively on MSI/MMR status may be overly simplistic as conventional grading based on the degree of glandular differentiation still holds greater prognostic significance in multivariate analysis.