Field test were conducted to select the best method for soil phosphorus test and to obtain the critical levels of each method for P fertilizer recommendations based on tobacco production. Truog method was recognized to be the mort suitable one for soil P test. N. Carolina, Bray No.1 and Olsen methods were followed in order. Bray No. 2 and Lancaster methods showed the least satisfaction. The critical levels of available P by lancaster, Bray No. 1, N. Carolina,01sen, and Truog methods were 55ppm, 32ppm, 21ppm, 19ppm, and 22ppm, respectively. Below the critical level a large yield response can be expected with 10kg $P_2O_5$/10a, above critical level little or no response is expected. In soils the more the initial available p, the more available P was carried over, while the more the total amount of silt and clay, the less the residual P was found. The increase in residual P by 10kg of $P_2O_5$ was greater than that of 5kg/10a. The chemical composition and the price of the cured leaf were not affected by the $P_2O_5$ application, indicating that the increase in the rate of $P_2O_5$ application would not exert on the quality of tobacco.