Multimodal biometrics systems combine different unimodal biometric sources to achieve high recognition accuracy. By overcoming the limitation of unimodal biometric systems in this paper, a multimodal biometric recognition system which combines two modalities, fingerprint and face based on score level fusion is proposed. For fingerprint trait, the features are built based on the minutiae points of ridge area, while Local Binary Patterns (LBP) is used for face trait. The procedures of pre-processing, features extraction and matching are proceeding independently to obtain the individual matching scores. Afterwards, the two calculated results are combined by a matching score level fusion scheme to make the final decision on declaring the person as genuine or an imposter