Groundwater is the sole source of drinking water for the population of Bainbridge Island. Increased use of groundwater supplies on Bainbridge Island as the population has grown over time has created concern about the quantity of water available and whether saltwater intrusion will occur as groundwater usage increases. A groundwater-flow model was developed to aid in the understanding of the groundwater system and the effects of groundwater development alternatives on the water resources of Bainbridge Island. Bainbridge Island is underlain by unconsolidated deposits of glacial and nonglacial origin. The surficial geologic units and the deposits at depth were differentiated into aquifers and confining units on the basis of areal extent and general water-bearing characteristics. Eleven principal hydrogeologic units are recognized in the study area and form the basis of the groundwater-flow model. A transient variable-density groundwater-flow model of Bainbridge Island and the surrounding area was developed to simulate current (2008) groundwater conditions. The model was calibrated to water levels measured during 2007 and 2008 using parameter estimation (PEST) to minimize the weighted differences or residuals between simulated and measured hydraulic head. The calibrated model was used to make some general observations of the groundwater system in 2008. Total flow through the groundwater system was about 31,000 acre-ft/ yr. The recharge to the groundwater system was from precipitation and septic-system returns. Groundwater flow to Bainbridge Island accounted for about 1,000 acre-ft/ yr or slightly more than 5 percent of the recharge amounts. Groundwater discharge was predominately to streams, lakes, springs, and seepage faces (16,000 acre-ft/yr) and directly to marine waters (10,000 acre-ft/yr). Total groundwater withdrawals in 2008 were slightly more than 6 percent (2,000 acre-ft/yr) of the total flow. The calibrated model was used to simulate predevelopment conditions, during which no groundwater pumping or secondary recharge occurred and currently developed land was covered by conifer forests. Simulated water levels in the uppermost aquifer generally were slightly higher at the end of 2008 than under predevelopment conditions, likely due to increased recharge from septic returns and reduced evapotranspiration losses due to conversion of land cover from forests to current conditions. Simulated changes in water levels for the extensively used sea-level aquifer were variable, although areas with declines between zero and 10 feet were common and generally can be traced to withdrawals from public-supply drinking wells. Simulated water-level declines in the deep (Fletcher Bay) aquifer between predevelopment and 2008 conditions ranged from about 10 feet in the northeast to about 25 feet on the western edge of the Island. These declines are related to groundwater withdrawals for public-supply purposes. The calibrated model also was used to simulate the possible effects of increased groundwater pumping and changes to recharge due to changes in land use and climactic conditions between 2008 and 2035 under minimal, expected, and maximum impact conditions. Drawdowns generally were small for most of the Island (less than 10 ft) for the minimal and expected impact scenarios, and were larger for the maximum impact scenario. No saltwater intrusion was evident in any scenario by the year 2035. The direction of flow in the deep Fletcher Bay aquifer was simulated to reverse direction from its predevelopment west to east direction to an east to west direction under the maximum impact scenario.