Characteristics of urban artificial landscape have permeated into natural and semi-natural landscapes. Prior studies focused on landscape pattern analysis and description in rapid urbanization areas but neglect the dynamic changes between structures and connectivity of artificial urban landscape and their internal interactions. We measured two landscape indices (landscape structures and landscape connectivity) and presented an integrated approach (the cointegration and causality) to explore the urban expansion of Dongguan, a rapidly growing city in southern China, since 1988 to 2009. Analyses of seven landscape indices indicates a clear dynamic change in the expansion of Donguan. We divided this expansion into four different periods, that is early natural and semi-natural landscapes dominated urban areas in 1988-1990, followed by rapid 'extension' and 'linear development' period through 1990-1999, and then urbanization developed steadily period ('sprawl') during 1999-2005, and the recent 'infilling' stage since 2005-2009 characterized by urban development? The results show that: (1) According to variation characteristics of seven indices, the period is divided into four stages. The first is 1988-1990, in which the natural and semi-natural landscapes such as farmland, woodland were dominant, while urban artificial areas were small and had low density and aggregation. The second period is 1990-1999, in which urban artificial landscape increased rapidly with high density and aggregation. Then, we named the urban land expansion as 'extension' and 'linear development' type. The skeleton of the city gradually formed and the urban artificial landscape was in the dominant position, while natural and semi-natural landscapes such as farmland, woodland were small with low density and aggregation. The third period is 1999-2005. The total area of urban artificial landscape grew steadily and we named the urban land expansion as 'sprawl' type. The last is 2005-2009. Urban artificial landscape grew further and we named the urban land expansion as 'infilling' type. (2) There exist cointegration relationships between class area (CA) and probability index of connectivity (PC), aggregation index (AI), contagion index (CONTAG), and connectivity indices IIC and PC. This indicates that there is a long-term equilibrium relationship among these variables and thereby no internal mechanism would break down the equilibrium of urban system. (3) Granger causality test shows that landscape connectivity indices (IIC and PC) are the Granger Causes of landscape structure indices (CA, AI, CONTAG), which suggests the landscape connectivity indices can be used to predict the change of landscape structures.