Huge increase in everything which operates on a network is called congestion. These things could be information packets in internet network or vehicles in road networks. Congestion can disorder the operation of complex networks which are traorting the information packets or particles e.g. in Internet. The first model which has been investigated in this thesis is a model that determines a particular destination for every information packet that is created in the system, and when it gets its destination, it will be vanished. The type of phase transition in this model is related to the routing rule of every node. Then we investigate this model on a 2-D lattice and hierarchical networks. Then another model has been investigated for congestion phenomena and diffusion of particles which is basically about hopping of particles from every node to its neighbours stochastically. The type of phase transition in this model is related to the structure of network and the probability of rejecting particles. A similar phase transitive behaviour has been observed like in the previous model. Basis of this model is defining two parameters q and which are respectively the probability of being empty and the probability of having number of particles greater than for the queue length of every node. Then this model is investigated on semi-bipartite network in which the generation of particles is only in peripheral nodes. Then the type of transition between different phases has been explained which is the origin of hysteresis in this process. Finally all of these predictions are verified by computational simulation. Key words: complex networks, congestion phenomena, phase transition.