Abstract Theoretical study, reported in this paper, qualifies unique mechanisms of water coning in gas wells. Water coning in gas wells has been understood as a phenomenon similar to that in the oil wells. It is shown, however that both the water inflow mechanism and its impact on well's productivity are substantially different. It is shown, for example, that, after water breakthrough, the oil-water interface at the well's completion would continue to cone, while the gas-water interface reverses at the top of the cone. Analyzed in the paper are the results of a conventional simulation of water coning in gas wells showing that water could affect productivity only at the very late stage of well's life. However, field data, shown in the paper, evidence early and severe water problems. This contradiction is explained in the paper by including the effects of Non-Darcy flow, perforation density and the ratio of vertical-to-horizontal permeability in modeling of water coning in gas wells. Results from numerical simulation combined with analytical models show that an early water breakthrough and a considerable increase in water production may result from combined effects of increased vertical permeability, lower density of perforation and high-velocity gas flow around the wells. Introduction Water coning in gas well has been understood as a phenomenon similar to that in oil well. In contrast to oil wells, relatively few studies has been reported an aspect of mechanisms of water coning in gas wells. Muskat 1 believed that physical mechanism of water coning in gas wells is identical to that for oil wells; moreover, he said that water coning would be less serious difficulties for wells producing from gas zone than for wells producing oil. Trimble and DeRose 2 supported Muskat theory with water coning data and simulation for Todhunters Lake Gas field. They calculated water-free production rate using Muskat-Wyckof 3 model for oil wells in conjunction with the graph presented by Arthurs 4 for coning in homogeneous oil sand. The results were comparing with a field study with a commercial numerical simulator showing that the rates calculated with Muskat-Wyckof 3 theory were 0.7 to 0.8 those of the coning model for a 1-year period. Kabir 5 used the analogy between high oil mobility well and a typical gas well, to investigate gas well performance coning water in bottom-water drive reservoir. He built a numerical simulator model for a gas-water system. He concluded that permeability and pay thickness are the most important variables governing coning phenomenon. Other variables such as penetration ratio, horizontal to vertical permeability, well spacing, producing rate, and the impermeable shale barrier have very little influence on both the water-gas ratio response and the ultimate recovery. McMullan and Bassioni (6) believed that water coning behaves differently in gas wells than in the oil wells. Using a commercial numerical simulator they got similar results than Kabir (5) for the insensitivity of ultimate gas recovery with variation of perforated interval and production rate. They demonstrated that a well in the bottom water-drive gas reservoir would produce with small water-gas ratio until nearly its entire completion interval is surrender by water. In this study, water problems begin when recovery factor is less than 30%. Fig. 1 shows water-gas ratio and gas recovery factor from field data of a gas well. It shows water production started after 404 days when the recovery was 22%. This well was killed for water production after 600 days of gas production when the recovery factor was 28%. Fig. 2 shows gas and water production rate versus time for another gas well. It shows water production star after 119 days of production. Gas rate was reduced from 6.0 MMSCFD to 4.0 MMSCFD due to water production rate of 30 BPPD. These two field data shows early water production in gas wells.