The complexity-entropy causal plane (CECP) has been widely discussed recently. It can measure the information of sequences from two perspectives to reflect their structural details. But in experiments we find that, as a method based on probability space, the original CECP is not sensitive to the shape of the probability distributions, which may lead to inaccurate structural feature differentiation. Therefore, we propose a novel normalized complexity-entropy causal plane based on the modified Fisher information measure (NF-CECP) to solve the problem. The modified Fisher information measure (MF) and divergence score (D-score) are two important parameters to characterize the signals from possibility space and divergence, respectively. According to simulation experiments, it could be intuitively found that NF-CECP can distinguish Gaussian white noises (GWN) from ARFIMA sequences, but the original CECP fails. We also apply this method to financial time series in different time periods. It reveals that the structural characteristics of financial time series change over time. And to a certain extent, the results reveal the growing economic ties between China and the United States.