The evolution of the industrial ecosystem is extremely important for industrial transformation and growth, and it is the frontier direction of industrial ecology. The industrial ecosystem is an analogue of a biological ecosystem and conducts material exchange and energy flow between different industries and the environment in a certain area. In such a system, the consumption of energy and materials is optimized, waste generation is minimized, and the effluents of one process serve as the raw material of another process. An industrial ecological network is a topological structural network consisting of the ecological relationships among different components, such as material or energetic relationships. In this paper, we construct a physical urban industrial ecological network based on an input-output analysis. Based on the ecological network analysis method, we suggest three characteristics indicators, roundput, eco-efficiency, and ascendency, to analyze the evolution of the urban industrial ecosystem. Roundput serves to describe the condition of an industrial ecosystem as it suits the purpose of being opposite to the term ‘throughput’, which is commonly used to illustrate the linear nature of the material flows in an industrial system; from raw materials, to products, to wastes. Ascendency, which combines system activity and organization, provides a single measure of the unitary process of growth and development. In addition, we use Beijing as an example to study the evolutionary tendency of the Beijing industrial ecosystem from 2005-2014. The following conclusions are drawn: (1) from tendency analysis, the three characteristics of the Beijing industrial ecosystem significantly increased over time. The evolution tendency of roundput and ascendency were nearly the same; the overall trend was upward with time and the numerical value in 2007 was significantly higher than that of other years. Eco-efficiency showed a trend of wave growth over the year. There were four periods of eco-efficiency growth based on the increasing speed. Period one was from year 2005 to 2007 (average increasing speed was 6.03% per year). Period two was from year 2007 to 2009 (average increasing speed was 17.42% per year). Period three was from 2009 to 2011 (average increasing speed was 3.58% per year) and period four was from 2011 to 2014 (average increasing speed is 16.76% per year). The eco-efficiency value in period one and three increased mildly, whereas in period two and four it increased sharply. (2) From indicator value analysis, the Finn cycling index of each year was between 0 and 1, which indicated that cycling throughflow of the Beijing industrial ecosystem was greater than straight throughflow. The eco-efficiency value increased 2.4 times in 10 years. Moreover, the eco-efficiency value in 2013 was slightly lower than 1, whereas in 2014 the eco-efficiency value was higher than 1, which indicated that the eco-efficiency in the Beijing industrial ecosystem reached the optimum state in 2014 from 2005 to 2014. The A/C value ranged from 0.187 to 0.256, lower than 0.401, indicating sustainable development capacity of the Beijing industrial ecosystem is not at the optimum state. Future research will focus on constructing compartmental industrial ecological networks and establishing industrial symbiosis networks between waste by-products of industrial sectors.