The tumor microenvironment has many important features that contribute to the resistance of cancer both to chemotherapy and to radiotherapy, and understanding it is a primary goal of current oncology. As is often the case in biology, its complexity is daunting, as it displays an abnormal vascularisation, gradients of nutrients and metabolites, hypoxia and a generally high acidity. Moreover, the chemical gradients exist in a complex feedback with local cellular adaptation, and the presence of normal cells of different types – for instance cells from the stroma and from the immune system. Physicists can contribute to the understanding with models, both numerical and computational. Although none of them can fully capture the complexity, they can be improved in successive steps, and provide expert guidance to hand pick the most relevant features of tumors, and potentially grasp just those elements that can lead to better cures. In this talk I shall describe the efforts undertaken be my group and by me, and I discuss their possible developments.