A series of bench scale coagulation experiments were carried out to test phosphorus removal from primary and secondary effluents of six regional municipal sewage treatment plants in the city of Kunming, Yunnan Province, China. Six commonly used metal salt coagulants (Aluminum chloride [AlCl3], aluminum sulfate [Al-2 (SO4)(3)], poly-aluminum-chloride [PAC], ferric chloride [FeCl3], ferric sulfate [Fe-2 (SO4)(3)], poly-ferric-sulfate [PFS]) were tested and compared on their phosphorus removal efficiencies. An empirical chemical phosphorus removal model was proposed and calibrated with the experimental data. The proposed model deals with the relationship between initial TP-concentration, residual TP-concentration, and metal salts dosage. The test results indicated that there was no significant difference between the tested metal salts in their phosphorus removal efficiencies. The efficiency of chemical phosphorus removal mainly relied on metal salt purity. This study showed a potential to use the empirical model to instruct chemical phosphorus removal in wastewater treatment plants for the high relativity (R=0.9189).