Recently, the utility of Sample Entropy(SampEn) as a complexity measure was shown via applying to time series data generated from in a variety of systems. However, online estimation method of SampEn index has not been developed yet. If SampEn can be estimated online, we can apply this index to time-varying system. In this paper, we developed the recursive SampEn algorithm to estimate the changes of system complexity online. In additon, we verified the utility of this algorithm by simulations. Consequently, we assure that this algorithm can be applied to time series generated from in a variety of time-varying system potentially.