Urban heat island (UHI) as a worldwide urban climate issue has long been studied. It was found as a phenomenon at first and its intensity were then quantified and even simulated. Nowadays people are more concerned about the mitigation of UHI effects. Suggested mitigation methods include (1) advocating energy conservation to reduce excessive anthropogenic heat emissions, (2) using appropriate building materials to construct so-called "cool roofs" or "green roofs" to alter urban landscape matrix and (3) reasonable urban planning and landscape design. With a still on-going propaganda of a 'low-carbon’ life style, the study of landscape matrix alteration has been well documented and applied. The changing of landscape pattern to mitigate UHI effects is however poorly carried out, partly due to theoretical and computational weakness. The problem with new building materials is that most of them are either expensive or need to renovate in a short period, so a good landscape planning and a better landscape design might be more constructive and efficient for mitigating UHI effects, especially in developing countries, where urbanization is still very fast. But a good landscape planning concerning UHI effects relief asks for a reasonable landscape pattern analysis with respect to UHI, and besides, appropriate configuration of the landscape matrix requires sound landscape pattern analysis, too, thus, it is important to carry out landscape pattern analysis for urban heat island mitigation. However to date, the landscape pattern analysis provides not much advices for landscape planning. In this paper, we first reviewed the research of the urban heat island history and related concepts briefly, such as Atmospheric UHI (AUHI) and Surface UHI (SUHI), and then focused on those studies mostly of the past decade dealing with UHI from the view of landscape pattern. Generally, two methods were used to analyze UHI effects; one is the traditional landscape metrics and statistical method-they are similar, the other is the simulation method. Different UHIs were treated because of data sources and mechanisms. SUHI were often related to landscape metrics and statistics of land surface temperature were made simultaneously, while meteorological data is used in AUHI analysis which is either coarse or fine in spatial resolution, so it was rarely related to landscape metrics. However, landscape metrics and statistical results are less helpful to landscape design of the unclear ecological meaning. Simulation methods provided convictive results, but they were scarcely done from the view of landscape pattern. As a result, less landscape pattern parameter was put into or derived from simulation model. The difficulties and shortcomings of these studies were discussed from viewpoint of the data source and the methods or the involvement of landscape pattern parameter. The future trends on urban heat island were finally proposed: (1) to collect or derive appropriate data based on the purpose of the study; (2) to conduct more landscape pattern simulation research rather than statistics, and furthermore to simulate 2 D (Dimensional) and 3 D and to construct 2/3D landscape indexes in UHI simulation models; (3) to use more multi-spectral and the thermal remote sensing data, and in-situ or simulated data as a verification to build mechanic meso-scale landscape model or index, in the purpose of evaluating or predicting the UHI effects in a meso-scale; (4) to integrate more socio-economic data and interdisciplinary methods during UHI analysis with landscape pattern.