The authors have been proposing a Reliability Based Design (RBD) scheme for practicing geotechnical engineers that is termed GRASP (Geotechnical Reliability Analysis by a Simplified Procedure). In the sister paper submitted to this conference by the same authors, a simplified method to evaluate the effects of soil spatial variability on the performance of the geotechnical structures is proposed. The method is based on the LA (local average) over the geotechnically appropriate volume. By focusing on the LA, it is possible to establish a clear-cut statistical estimation theory which gives the bases for evaluating statistical estimation error. In this paper, the outline of GRASP is summarized in the first half of this paper with emphasis on the way to evaluate the statistical estimation error. The latter half of the paper is devoted to the consideration on the major uncertainty sources in geotechnical design based on results of analyses by GRASP on various typical geotechnical structures, such as piles, shallow foundations and embankments on soft ground.