A method for fraud modeling based on shared computing device usage includes: storing transaction data entries, each including a transaction date and/or time, account identifier, and device identifier associated with a computing device; receiving a transaction message for a payment transaction, the transaction message including a specific device identifier, primary account number, and additional transaction data; identifying transaction data entries where the included device identifier corresponds to the specific device identifier; determining a fraud risk rating based on a number of unique account identifiers included in the identified transaction data entries over a predetermined period of time; and transmitting the transaction message and the determined fraud risk rating to a financial institution associated with the primary account number.