The facts of early non-performance of iron-based underground water pipes of the Kathmandu Valley (Nepal) are mostly due to the corrosive nature of the surrounding soil factors. A comparatively high percentage of the damage to the water pipelines in Nepal is anticipated, compared with the data reported in the developed countries. For the reasons, the quantitative data analysis of six kinds of the corrosive soil factors (i.e., moisture-holding capacity, pH, electrochemical resistivity, redox potential, Cl−, and SO4
2− ions) of 15 specimens, sampled from the Manohara Town Planning (MTP) area, was carried out using a novel probabilistic model, which is based on the internationally accepted standards (i.e., ASTM, NACE, AASTHO) with some modifications. The corrosivity of each soil factor was categorized into four corrosive groups (CGs) based on the conventionally applied classification methods. The experimentally obtained data of the soil factors were then used for proximate analysis using a new non-deterministic (probabilistic) model to classify the corrosive sub-groups (CSGs) of the soil specimens. This model would be a meritorious method to assess the soil corrosion condition of any area of the country, where the engineers and technologists could be applied pertinent methods or materials for the buried-pipe works in the future.