Oil wells that intersect a potential CO2 storage zone in a depleted oil and gas field may provide leakage pathways. It is essential to estimate the field‐scale leakage risk associated with these wells. In this study, a risk‐based approach is used to estimate the risk of leakage. Existing reduced‐order models for well leakage are used to quantitatively estimate well leakage rates for cased‐cemented, cased‐uncemented, and open wellbores. For each existing well that intersects the storage zone, we introduce the well leakage index (WLI), which accounts for wellbore geometry, distance from the injection well, buffer layers between the storage zone and underground sources of drinking water, and the nature of storage zone boundary type. For an initial injection well location, the total site well leakage index (SWLI) is calculated, which is the summation of the WLI for all of the wells. Next, the injector location is varied areally and SWLI is calculated for a specified number of potential injector well locations in the storage zone area. Small values for the SWLI correspond to low well leakage potential, indicating where injection well locations can be considered. The developed criterion provides a means to systemically find the areas with highest and lowest well leakage potential for a storage zone. Due to the reduced order nature of the developed method, it should be a useful tool in the planning and execution phase of the CO2 geological sequestration process. © 2019 Society of Chemical Industry and John Wiley & Sons, Ltd.