In this paper, we investigate adaptive settings for the initial interpolation vector in our previous iterative time-domain compression method for multiple-input multiple-output (MIMO) channel state information (CSI). The iterative MIMO-CSI compression method is based on time-domain CSI compression with frequency-domain CSI bundling and interpolation for all combinations of transmitter and receiver antennas. Previously, the method for generating the initial interpolation vector, which reuses the actual frequency-domain CSI variation, was proposed in order to take into account the periodical channel variation in the frequency domain. However, in this method, the appropriate length interpolation vector is dependent on the delay spread of the channel. In order to address this problem, this paper proposes adaptively controlling of the length of the interpolation vector based on the correlation of the CSI vector in the frequency domain. The numerical results show that the proposed method further improves the accuracy of the compressed CSI or enhances the CSI compression capability of the previous method, which assumes a fixed length for the interpolation vector.