Abstract: This study was designed to determine endogenous fecal calcium and phosphorus losses and true fecal calcium and phosphorus digestibility with soybean meal for growing barrows by the Linear Regression Analysis Technique (REG). Six Yorkshire×Landrace barrows, with averaged initial body weight of （28.6±1.76） kg, were fed in the metabolic cage. According to a 6×6 Latin square design, six soybeancornstarch based diets, containing six levels of calcium (0.04%, 0.09%, 0.13%, 0.17%, 0.22% and 0.26%) and phosphorus (0.09%,0.18%,0.27%,0.35%,0.44% and 0.53%), were formulated in six experimental periods. Each experimental period comprised 8 d with 6 d adaptation and 2 d collection of fecal samples. When expressed as g/kg DM diet intake,increasing linear relationships（P=0.002<0.01）were found between fecal outputs and dietary inputs of Ca. The endogenous Ca output of barrow was calculated as 0.622 5 g/kg DMI and true Ca digestibility was 44.34% when the dietary inputs of Ca point to zero on the regression curve.When expressed as g/kg DMI, increasing linear relationships（P=0.025<0.05）were found between fecal outputs and dietary inputs of P. The endogenous P output was calculated as 1.077 1 g/kg DMI and true P digestibility was 48.78% when the dietary inputs of P point to zero on the regression curve. These results suggested that in the soybean diets of different Ca and P levels, apparent digestibility of Ca and P vary a lot,but true digestibility of Ca and P are relatively stable. Apparent digestibility of Ca and P were lower than true digestibility by 41.50% and 35.51%, respectively. Current diet formulation using apparent Ca and P digestibility leads to underestimation of Ca and P biological availability and excessive intake and excretion Ca and P in pigs. True Ca and P digestibility should be determined and used in diet formulation for pigs.