Based on the principle of information spread，normal information spread estimation method(NISEM) is proposed to generate probability density functions of geotechnical parameters with small samples. The new method is distribution-free for no classical theoretical distributions were assumed in advance；and the inference result provides a universal form of probability curves. Then it presented a good solid mathematical and physical performance. Kolmogorov-Smirnov test is used to verify the availability of this method. The results of case study show that the index Dn in NISEM is all smaller than those in the classical finite comparative method(FCM)；and the NISPM function is more accurate and effective than FCM. It shows that the presented method can be used in stochastic reliability analysis of geotechnical engineering.