The financial crises that have shaken global economy have largely been the result of a discretionary influence abuse of managers and administrators, as well as a consequence of the lack of independence of the auditors, to the detriment of other subjects. This paper aims at setting the foundations and testing a deterministic model of fraud detection, following a study made on the financial statements belonging to a sample of 65 quoted organizations, guided by the classification of financial fraud made by the international standards on auditing. The financial auditor resorts to diagnosis analysis in order to identify, using specific indicators, errors, inconsistencies, accounting manipulations that may represent a basis of fraud. This has lead to testing, within this paper, the existence of interdependencies between a series of ratios (independent variables) and the financial frauds that may occur within a company. Therefore, the most representative ratios have been selected, for building a mathematical model (score function) that would provide the financial auditor with information necessary for fraud detection.