Preloading without a surcharge is a common method for ground improvement. Thereare however uncertainties related to the number of site investigations and the partialfactor method has been identified as a problem. This thesis proposes a probabilisticdesign approach, using a Monte Carlo simulation to calculate the failure probabilityin the serviceability limit state for preloading without a surcharge. The methodwas applied to a case where the possibility of using preloading without a surchargewas determined. A parameter influence and sensitivity analysis were performed todetermine what parameters were most important for the calculation. Problems withthe generation of random samples for the parameters; preconsolidation pressure andlimit pressure were identified, and four different methods of generating the randomsamples were tested, and discussed.The failure probability was calculated as a function of preloading time which wasused to determine what preloading time is needed to fall below the acceptable failureprobability of 5%. The required preloading time was found to be 580 days. Themost important parameters were found to be preconsolidation pressure, the modulusM0 and the coefficient of vertical consolidation.The proposed method is working and has several advantages, among them are theability to calculate the failure probability and the compatibility with the observationalmethod. The model uncertainty has been discussed, and a general commentis that with more research can the model uncertainty be decreased. Only 1D consolidationis considered in the method, this simplification is very practical as 2D and3D effects often can be neglected.