We discuss our recent theoretical and experimental work in cardiac systems. The first is about the rhythmic variation as the cardiac myocyte cell culture synchronized, and how it can be explained in terms of the frequency enhancement phenomenon in coupled excitable systems. Another work concerns the prediction of self-terminating ventricular fibrillations (VF) based on the analysis of time-series signals from ventricles and near the sinus atrial node by the cross-wavelet power spectrum and cross-Fourier power spectrum methods. The success rate of our prediction criteria is about 80-90%. Our findings suggest that a heart under VF can recover its sinus rhythm only when the sino-atrial node of the heart is not under strong influence of the VF from its ventricle. Finally the dramatic reduction of cardiac alternans by small perturbations in pacing scheme. Alternans are alternating long and short action potential durations in heart beats, which is a precursor of the fatal. Predictions and validity of this control method have been verified by both experiments performed with isolated heart preparations and numerical simulations. A nonlinear return map for this novel pacing scheme based on action potential duration restitution response is proposed to explain the working mechanism of the control. Extending this novel control method to general complex dynamics can be achieved.