Detection of an optimal panel of biomarkers capable of predicting a patient's risk of major adverse cardiac events (MACE) is of clinical significance. Due to the high dynamic range of the protein concentration in human blood, applying proteomics techniques for protein profiling can generate large arrays of data for development of optimized clinical biomarker panels. The objective of this study is to discover a panel of biomarkers for predicting risk of MACE in subjects reliably. The development of immunoassay can only tolerate the complexity of the prediction model with less than ten selected biomarkers. Hence, traditional optimization methods, such as genetic algorithm, cannot be used to derive a solution in such a high-dimensional space. In this paper, we propose an improved genetic algorithm with the local floating searching technique to discover a subset of biomarkers with improved prognostic values for prediction of MACE. The proposed method has been compared with standard genetic algorithm and other feature selection approaches based on the MACE prediction experiments.