For the uncertain dynamic decision-making problem that the decision information is interval number, under the condition that the attribute weights and time weights are unknown, a dynamic multiple attribute decision-making model considering both decision information and decision preference based on improved vector similarity is proposed. Based on the relative similarity relations of the interval decision information and the importance of attribute, the relative similarity degree minimum programming model is constructed to determine the index weights. Taking the time value of decision information and the decision maker's preference into account, the maximum entropy model is established to determine the time weights. Combining the shortcoming of vector similarity calculation, a measure method of vector similarity based on vector projection idea is given, so that the uncertain dynamic decision-making model is constrated. Finally, an example is given to illustrate the rationality and effectiveness of the proposed model.