Chemical treatment of non-point derived pollution often suffers from undesirable oscillations in purification efficiency due to variations in runoff water quality. This study examined the response of the chemical purification process to variations in water quality using a 2^k factorial design for runoff water rich in humic substances. The four k factors evaluated and the levels applied were: organic matter as dissolved organic carbon (DOC) (20–70 mg/L), suspended solids (SS) (10–60 mg/L), initial water pH (4.5–7), and applied coagulant dosage (ferric sulphate) (35–100 mg/L). Indicators of purification efficiency were residual concentration of DOC, SS and total phosphorus (tot-P). Analysis of variance and factor effect calculations showed that the initial DOC concentration in raw water samples and its interactions with the coagulant dosage applied exerted the most significant influence on the chemical purification process, substantially affecting the residual concentration of DOC, SS and tot-P. The variations applied to the factors SS and pH only slightly affected purification efficiency. The results can be used in the design of purification systems with high organic matter load variation, e.g. peat extraction runoff.