针对人民币汇率收益率时间序列数据存在的跳变特征,采用跳扩散模型对其时间序列数据进行描述.为识别跳变规律并解决模型的参数估计问题,提出了基于跳辨识-MCMC的组合算法:即结合Lee-Mykland的跳辨识方法与MCMC(蒙特卡罗马尔可夫链)方法形成组合算法,利用仿真实验,通过误差分析得出组合算法在跳扩散模型参数估计方面效果明显优于单一MCMC方法.以人民币/美元日汇率数据为样本进行实证分析,结果表明组合算法不但能较为准确地识别出汇率收益率的跳变时刻及规律,而且其模型参数估计的有效性大大提高. In order to identify the transition law and solve the parameter estimation problem of the model, the jump diffusion model is used to describe the transition characteristics of the time series data of RMB exchange rate returns, and the combinational algorithm based on jump identification and MCMC is proposed: That is to say, combining with Lee-Mykland’s jump identification method and MCMC (Monte Carlo Markov chain) method, a combination algorithm is formed. By using the simulation experiment, the error analysis shows that the combined algorithm is obviously better than the single MCMC method in jump diffusion model parameter estimation Taking the RMB / USD daily exchange rate data as an example, the result shows that the combination algorithm not only can accurately identify the transition time and rule of the exchange rate return, but also greatly improves the validity of the model parameter estimation.