In parameter estimation, geotechnical engineers are accustomed to provide their best estimates with engineering judgments. But the bounds or indication of the significance of their estimates are seldom given. It is difficult to routinely implement reliability-based design if gauging uncertainties of engineers' estimates is not mandatory or does not become a usual practice. The objective of this paper is to advocate the importance of gauging uncertainty in the course of parameter estimation. A framework for parameter estimation and uncertainty analysis incorporating engineering judgment is illustrated based on Bayesian inversion. The data uncertainty is considered by additive noise model and the degree of confidence of engineering judgment is accounted for by the prior information in the Bayesian paradigm. The information contained in the posterior distribution resulting from the Bayes theorem is interpreted by optimal estimator, uncertainty of estimation, and resolution operator. The advantages of Bayesian inversion over commonly-used least square method are illustrated by some simple examples.