Certain aspects of the present disclosure provide techniques for predicting a transaction time based on user position data. In certain aspects, a method for predicting a transaction time based on user position data includes obtaining a transaction record and one or more user positions associated with a user. The method also includes obtaining one or more business records associated with each respective user position. The method further includes calculating one or more similarity scores, where each similarity score is based on a similarity between a respective business record and the transaction record. The method also includes associating the transaction record with a business record based on a maximum similarity score of the one or more similarity scores. The method further includes determining a predicted transaction time for the transaction record based on at least a timestamp of a user position associated with the business record associated with the transaction record.