The links between pairs of nodes within many real-world networks change over time. Thus, there has been a recent boom in studying temporal graphs. Recognizing patterns in temporal graphs requires a similarity measure to compare different temporal graphs. To this end, we propose to study dynamic time warping on temporal graphs. We define the dynamic temporal graph warping distance (dtgw) to determine the (dis)similarity of two temporal graphs. Our novel measure is flexible and can be applied in various application domains. We show that computing the dtgw-distance is a challenging (in general NP-hard) optimization problem and identify some polynomial-time solvable special cases. Moreover, we develop a quadratic programming formulation and an efficient heuristic. In experiments on real-word data we show that the heuristic performs very well and that our approach performs favorably in de-anonymizing networks compared to other approaches. networks, traffic networks, attack networks in computer security, or protein-protein-interaction networks in biology [10,16,18,23]. Many processes described by temporal graphs naturally vary in duration and temporal dynamics (for example, chemical reactions might proceed with different speed), which makes data mining tasks such as classification challenging. Hence, in order to perform classification or clustering on temporal networks, one needs to find suitable (dis)similarity measures, which has seemingly not been done so far.Our paper proposes such a measure. We introduce a novel (dis)similarity measure on temporal graphs based on vertex signature graph distance and dynamic time warping, called dynamic temporal graph warping (dtgw). Dynamic time warping allows to filter out variations in temporal dynamics. Thus, by combining established methods from graph-based pattern recognition and time series data mining in a nontrivial way, we obtain a suitable tool to analyze temporal network data. We study its computational complexity, develop efficient algorithms and study their behavior on real-world data, the latter clearly indicating the strong potential for future applications.