Bridge networks are expensive and complex infrastructures and are essential components of today's transportation systems. Despite the advancement in computer aided modeling and increasing the computational power which is increasing the accessibility for developing the fragility curves of bridges, the complexity of the problem and uncertainties involved in fragility analysis of the bridge structures in addition to difficulties in validating the results obtained from the analysis requires precaution in utilization of the results as a decision making tool. The main focus of this research is to address, study and treatment of uncertainties incorporated in various steps of performance based assessments (PBA) of the bridge structures. In this research the uncertainties is divided into three main categories. First, the uncertainties that come from ground motions time and frequency content alteration because of scarcity of the recorded ground motions in the database. Second, uncertainties associated in the modeling and simulation procedure of PBA, and third uncertainties originated from simplistic approach and methods utilized in the conventional procedure of PBA of the structures. Legitimacy of the scaling of ground motions is studied using the response of several simple nonlinear systems to amplitude scaled ground motions suites. Bias in the response obtained compared to unscaled records for both as recorded and synthetic ground motions. Results from this section of the research show the amount of the bias is considerable and can significantly affect the outcome of PBA. The origin of the bias is investigated and consequently a new metric is proposed to predict the bias induced by ground motion scaling without nonlinear analysis. Results demonstrate that utilizing the predictor as a scaling parameter can significantly reduce the bias for various nonlinear structures. Therefore utilizing the new metric as the intensity measuring parameter of the ground motions is recommended in PBA. To address the uncertainties associated in the modeling and simulation, MSSS concrete girder bridge class were selected due to the frequency of the construction in USCS region and lack of seismic detailing. A large scale parameters screening study is performed using Placket-Burman experimental design that considers a more complete group of parameters to decrease the computational expense of probabilistic study of the structure's seismic response. Fragility analysis for MSSS bridge is performed and the effect of removing the lesser important parameters the probabilistic demand model was investigated. This study reveals parameters reduction based on screening study techniques can be utilized to increase efficiency in fragility analysis procedure without compromising the accuracy of the outcome. The results from this study also provides more direct information on parameter reduction for PBA as well as provide insight into where future investments into higher fidelity finite element and constitutive models should be targeted. Conventional simplistic PBA approach does not account for the fundamental correlation between demand and capacity models. A more comprehensive PBA approach is presented and fragility analysis is performed with implementation of a new formulation in the component fragility analysis for MSSS bridge class and the outcome is compared with the one from conventional procedure. The results shows the correlation between demand and capacity affects the outcome of PBA and the fragility functions variation is not negligible. Therefore using the presented approach is necessary when accuracy is needed.