It is increasingly recognized that global change research requires methods and strategies for combing process models and data in systematic ways. This is leading to research towards the application of model-data fusion approach. The model-data fusion is a new quantitative approach to model analysis and data assimilation that provides a high level of empirical constraint over model predictions based on observations. Applications of model-data fusion require (a) a model that describes the underlying physical, chemical and biological processes, (b) experimental observations and (c) an optimization tool. The optimization tool is used to find optimal estimates of model parameters or states by minimizing the differences between model predictions and experimental observations. Finding the optimal parameters can help us improve predictions or test alternative hypotheses embedded in the models. Model-data fusion can be used in several different ways: to estimate parameter values or in a sensitivity study that can be used to identify the observations required to estimate model parameters or to test our hypotheses. In this paper, we will review recent applications of model-data fusion in global ecology and paleoecology studies and highlight current progress and issues, potential problems and future challenges.