This thesis presents techniques for the modeling and identification of linear state space models of structural response to stochastic dynamic input. Two approaches to stochastic system identification are investigated and further developed within the state space modeling framework. These methods require only measurements of the response of the system for estimation of modal parameters.