Guaranteed solutions to the current transaction security challenges facing banking industries may not be possible on a large scale, if the solutions are not developed to prevent third party transaction in the used software. Thus, this research lends its voice in support of the move to develop a biometric platform for preventing third party transaction in the banking system. It is meant to use both human face and fingerprint in securing payment platform in the banking software. Hence, the study is motivated by the need to develop a payment platform that is self-secured and be able to prevent third party transaction. In carrying out this research, an in-dept analysis of the existing banking software (Finacle 10.8) used in United Bank for Africa (UBA) was carried out to ascertain its existing security features, also the existing up-to-date banking security models were reviewed in order to establish the gap filled by this research. The data for this research were collected through the use of Key Informant Interview Method (KIIM), site visit and system observation techniques. It was seen that the existing software does not prevent third party transaction and made the software to be secured for customer but not secure for people entrusted with the software (staff). Hence we proposed the used of human face and fingerprint to secure the payment platform in this software as against the use of account number. In designing the platform, the Unified Modeling language such as Use-Case, Class diagrams etc were used as the modeling tool. Similarly, the platform was developed using Object Oriented Analysis and Design Methodology (OOADM) with Rapid Unified Process (RUP) model to manage the software processes. In examining the designed architecture, the Open Group Architectural Framework and the Design Science approach were used to evaluate the architecture/design and the control mechanism framework. During the implementation of the platform, Microsoft Visual Studio 2008 was used as the Integrated Development Environment (IDE) and Microsoft SQL Server 2008 was used as the backend. Similarly, Visual C#.NET programming language was used to program the application control mechanism and ASP.NET was used to develop the application interfaces. The implementation of the software and the testing shows that the platform could secure the banking system using both human face and fingerprint biometric. This thereby makes the software to be able to prevent third party transaction. Similarly, this platform was subjected to load performance testing using Jmeter performance testing package in order to ascertain the system performance, that is, tail tolerance and scalability. The result captured from the use of the Jmeter was graphically presented using excel. The result shows that the tail tolerance of the platform build using our architecture is significantly better than its equivalent. Specifically, we established that our biometric platform is better than the current platform used in securing banking transaction and can prevent third party transaction.