Phosphorus(P) is an important macronutrient for plant but can also cause potential environmental risk. In this paper, we studied the long-term fertilizer experiment(started 1980) to assess the soil P dynamic, balance, critical P value and the crop yield response in Taihu Lake region, China. To avoid the effect of nitrogen(N) and potassium(K), only the following treatments were chosen for subsequent discussion, including: C0(control treatment without any fertilizer or organic manure), CNK treatment(mineral N and K only), CNPK(balanced fertilization with mineral N, P and K), MNK(integrated organic manure and mineral N and K), and MNPK(organic manure plus balanced fertilization). The results revealed that the response of wheat yield was more sensitive than rice, and no significant differences of crop yield had been detected among MNK, CNPK and MNPK until 2013. Dynamic and balance of soil total P(TP) and Olsen-P showed soil TP pool was enlarged significantly over consistent fertilization. However, the diminishing marginal utility of soil Olsen-P was also found, indicating that high-level P application in the present condition could not increase soil Olsen-P contents anymore. Linear-linear and Mitscherlich models were used to estimate the critical value of Olsen-P for crops. The average critical P value for rice and wheat was 3.40 and 4.08 mg kg~(–1), respectively. The smaller critical P value than in uplands indicated a stronger ability of P supply for crops in this paddy soil. We concluded that no more mineral P should be applied in rice-wheat system in Taihu Lake region if soil Olsen-P is higher than the critical P value. The agricultural technique and management referring to activate the plant-available P pool are also considerable, such as integrated use of low-P organic manure with mineral N and K. Phosphorus (P) is an important macronutrient for plant but can also cause potential environmental risk. In this paper, we studied the long-term fertilizer experiment (started 1980) to assess the soil P dynamic, balance, critical P value and the crop yield response in Taihu Lake region, China. To avoid the effect of nitrogen (N) and potassium (K), only the following treatments were chosen for subsequent discussion, including: C0 (control treatment without any fertilizer or organic manure), CNK treatment ( mineral N and K only, CNPK (balanced fertilization with mineral N, P and K), MNK (integrated organic manure and mineral N and K), and MNPK (organic manure plus balanced fertilization). The results revealed that the response of wheat Yield was more sensitive than rice, and no significant differences of crop yield had been detected among MNK, CNPK and MNPK until 2013. Dynamic and balance of soil total P (TP) and Olsen-P showed soil TP pool was enlarged significantly over consistent fertilization . However, the diminishing marginal utility of soil Olsen-P was also found, indicating that high-level P application in the present condition could not increase soil Olsen-P contents anymore. Linear-linear and Mitscherlich models were used to estimate the critical value of Olsen-P for crops. The average critical P value for rice and wheat was 3.40 and 4.08 mg kg -1, respectively. The smaller critical P value than in uplands indicates a stronger ability of P supply for crops in this paddy soil We concluded that no more mineral P should be applied in rice-wheat system in Taihu Lake region if soil Olsen-P is higher than the critical P value. The agricultural technique and management referring to activate the plant-available P pool are also available , such as integrated use of low-P organic manure with mineral N and K.