Drainage ditches used to remove surface runoff or connected to subsurface tile and mole drains may act as a major conduit of phosphorus (P) from agricultural lands to receiving waters. The extent of P transport via drainage ditches is potentially governed by the P status and retention characteristics of drainage ditch sediments. Twenty-six surface (0–5cm) and subsurface (5–15cm) sediment cores (10cm diameter) from 26 drainage ditches in four major New Zealand pastoral catchments were therefore characterised for P fractions and P retention capacities. Both surface and subsurface sediments were found to contain a significant amount of P and possess a range of P retention capacities. Phosphorus retention capacities in surface and subsurface sediments ranged from 2467 to 4197 and 2225 to 3891mgPkg⁻¹ sediment, respectively. They were significantly correlated (r=0.638–0.918; P≤0.001) with sediment chemical characteristics (pH, organic matter, and oxalate-extractable Al and Fe). Approximately 42–57% (±S.E.M. of 1.1–2.8) of P in drainage sediments was present as loosely bound fractions (non-occluded Al/Fe-P and carbonate-bound P), suggesting that drainage sediments may temporarily store P originating from agricultural catchments, and that P held in this storage pool may be readily released into the overlying drainage water. Calcium-bound P and occluded Al/Fe-P represented minor fractions in the drainage sediments, accounting for less than 10% of total P in the sediments (2–7 and 5–9% of total P, respectively). Drainage sediments also acted as a long-term P sink, since residual P represented a substantial fraction (6–39% of total P) in some drainage sediments. Drainage water contained not only soluble P but also particulate and dissolved organic P fractions, indicating that drainage management for P pollution control needs to consider all these three P fractions. Soluble P concentrations (0.006–0.019mgPl⁻¹) in drainage waters were not significantly correlated with loosely bound sediment P, suggesting a non-equilibrium status between sediment and overlying drainage water P. Although the surveyed sediments had high P retention capacities (44–84% of added 5000mgPkg⁻¹ sediment), up to 64–68% of P sorption sites on sediments were saturated with P, attributed to a long history of P loadings from agricultural runoff and subsurface flows. The results suggest that water-extractable or Olsen-extractable sediment P should be used in conjunction with the sediment P saturation index (degrees of P saturation) to identify drainage ditches that act as a potential source of P to receiving waters. Information obtained could then be used in targeting appropriate management practices to minimise P release from these ditches.