Background: Studies on the association between serum vitamin B6 status and colorectal cancer prognosis are limited and have yielded inconsistent results. This study investigated the association of pyridoxal 5′-phosphate (PLP) and pyridoxic acid ratio (PAr) index with colorectal cancer survival. Methods: A total of 1286 colorectal cancer patients diagnosed since 2010 were selected from the Guangdong Colorectal Cancer Cohort study. Serum levels of PLP, pyridoxal, and 4-pyridoxic acid were measured using ultra-high-performance liquid chromatography–tandem mass spectrometry. The study followed overall mortality and colorectal cancer-specific mortality until December 2023. Multivariable Cox proportional hazards regression models were applied to calculate hazard ratios (HRs) and 95% confidence intervals (95% CIs). Restricted cubic spline and stratified analysis were performed. Results: During a median follow-up of 77.36 months, 331 deaths were recorded, with 293 specifically attributed to colorectal cancer. Higher PLP levels were associated with a longer overall survival (HRQ4 vs. Q1, 0.63; 95% CI: 0.46, 0.87; p for trend = 0.008) and colorectal cancer-specific survival (HRQ4 vs. Q1, 0.62; 95% CI: 0.44, 0.87; p for trend = 0.006). Non-linear associations were observed between serum PLP and overall and colorectal cancer-specific survival (p for non-linear < 0.05). However, PAr was not significantly associated with either overall survival (HRQ4 vs. Q1, 1.03; 95% CI: 0.75, 1.41) or colorectal cancer-specific survival (HRQ4 vs. Q1, 1.01; 95% CI: 0.72, 1.42). The association between serum PLP and both overall survival and colorectal cancer-specific survival (p for interaction < 0.05) varied by alcohol drinking status. Conclusions: Higher serum PLP levels, but not PAr, may be associated with improved overall and colorectal cancer-specific survival.