Soils are increasingly acknowledged as a source of P discharges in surface water. In order to monitor the transfer processes from soil to streams and rivers, a simple, sensitive, accurate, and robust method is required to determine trace concentrations of total dissolved P (TDP) in soil solution. This matrix provides a unique challenge to the analyst because it often contains trace concentrations of P in the presence of higher concentrations of organic and colloidal fractions that can interfere. Leachate waters were collected from lysimeter experiments covering 10 different soil types from the UK, ranging from sandy lowland agricultural grassland soil to an upland natural peaty gley soil, to critically evaluate the performance of different methods of analysis. Solutions of model compounds also provided measures of performance. Three methods involving digestion and colorimetric determination and one direct instrumental method were assessed. Methods evaluated were a mild digestion (persulfate), two rigorous oxidation procedures (peroxide‐Kjeldahl and nitric acid‐sulfuric acid) and inductively coupled plasma‐optical emission spectrometry (ICP‐OES). The ICP‐OES lacked the sensitivity to determine P concentrations below 100 µg L−1. Nitric acid‐sulfuric acid, a multi‐staged procedure, gave erratic recovery and was vulnerable to contamination. The best method for the analysis of soil solution or leachates was the acidified persulfate digestion because it produced reliable and accurate data. Thus we recommend that acid persulfate digestion is used to determine TDP in soil solution.