A new approach is proposed for the simulation of interacting human physiological subsystems. The new simulation environment will allow users to build, analyze and solve coupled physiological system models based on available distributed subsystem models. The subsystem models as an abstract form of physiological knowledge are shared through the Internet. The simulation knowledge network includes a distributed model database, an information server, and an analysis and solution environment that can be downloaded from the server. To represent the coupled system, we use Differential Algebraic Equations (DAE) as the basic format. There are three common sources of algebraic constraints in models of physiological systems. They are causal conflict, simulation optimization, and multistage processes. The singularly perturbed sliding manifolds method is chosen for solving DAEs. Possible future improvements are discussed. Examples of simulation of coupled circulatory, thermoregulatory, and renal systems are presented to demonstrate the simulation environment.