Host-directed therapies (HDT) that modulate host-pathogen interaction offer an innovative strategy to combat Mycobacterium tuberculosis (Mtb) infections. When combined with conventional anti-tuberculosis regimens, HDT strategies could contribute to improving treatment outcomes, reducing treatment duration, and preventing resistance development. It is however challenging to evaluate the interplay of host-pathogen interaction events in response to HDT strategies, and to translate experimental findings towards patients. Quantitative understanding of the multi-faceted nature of the host-pathogen interactions is vital to rationally design HDT strategies. Here, we (1) provide an overview of key host-pathogen interactions as basis for HDT strategies, (2) discuss experimental models to characterize host-pathogen interactions relevant for HDTs, and (3) discuss the utility and approaches of quantitative systems pharmacology (QSP) models to inform design of HDT strategies against Mtb infections. QSP models can be used to identify and optimize treatment targets, to facilitate preclinical to humans translation, and to design combination treatment strategies.