This study compounds three types of coagulants (AlCl₃, FeCl₃, Fe₂(SO₄)₃) with promising effects on TP removal of domestic sewage. The optimum conditions for TP removal using1compounded coagulants are determined by Plackett-Burman (P-B) design, steepest ascent, and Box-Behnken design. The adequacy of the quadratic regression model is evaluated by analysis of variance (ANOVA). Results show that initial pH, AlCl₃, and Fe₂(SO₄)₃ are the significant factors for TP removal. F-test, P-value (Prob>F), coefficients R², coefficient of variation, and adequate precision analysis demonstrated the goodness of fit for the regression model. The optimized conditions for TP removal determined by the response surface methodology are initial pH 5.2, AlCl₃ 45 mg/L, and Fe₂(SO₄)₃ 51 mg/L, respectively. The experimental TP removal efficiency (82.89%) agrees with the predicted response value (81.99%), thereby validating the feasibility of the model. Compared to single coagulants (AlCl₃, FeCl₃, Fe₂(SO₄)₃), the compounded coagulants demonstrate 3.29%, 7.59%, and 15.19% higher for TP removal; and 10.1%, 3.0%, and 10.3% higher for COD_Cr removal. In addition, the compounded coagulants also alleviate the potential hazards to human health due to the dosage decrease of aluminium salt coagulants.