Abstract: Diaphragm-type pressure cells are often used in geotechnical model tests and in-situ tests. The soil pressures are obtained through the measurement of the diaphragm deflection. With a gauge inserted into the soil or sand, the stiffnesses of the cells and the geo-medium are different. The soil arching in the soil or sand above the cells caused by the differential deformation will reduce the pressure on the diaphragm. In addition, different stress histories in the model tests result in different levels of influence. In the piled embankment multiple trapdoors model tests, the cells on the pile (fixed beams) experience a loading process, while the cells in the soil (movable beams) experience an unloading process. One loading calibration test and four unloading calibration tests are carried out on the same sand under the same relative density as in the model tests. The calibration results show that the stress-strain loading curves are almost linear and the unloading curves are quite similar to exponential ones. The calibration coefficients are obtained and then used to deal with the measurements. During the sand filling procedure, the average pressures obtained using the calibration coefficients coincide well with the self-weight of the embankments. During the settling (trapdoor movement) procedure, the pressures of the cells on the pile and in the soil are calculated using loading coefficients and unloading coefficients respectively. The results show that the soil pressure curves coincide well with those in the real model tests.