The principal aim of this research study was to improve the reliability of structural performance prediction for pavement structural condition. The intention of this research was to provide accurate information upon which decisions can be made, regarding when and how a particular road network needs to be rehabilitated, and where the associated budget should be allocated. The project focused on developing structural deterioration models for flexible pavements for the Queensland road network using Traffic Speed Deflectometer (TSD) data collected over a five-years period by the Department of Transport and Main Roads (TMR), Queensland. The Falling Weight Deflectometer (FWD) is a non-destructive testing device which has been used by many road agencies worldwide since 1980's, primarily for structural evaluation of flexible pavements for pavement design and evaluation of structural integrity. In early 2000, the world's first Doppler based TSD was developed by Greenwood Engineering. Currently more than ten TSD vehicles are in use worldwide. Two main advantages of the TSD technology are its ability to carry out continuous bearing capacity measurements, rather than measurements at discrete points, and the ability to conduct tests at traffic speed without the need for lane closure, as is the case for a stationary device, such as, the FWD. To achieve the desired outcomes, of this study a comprehensive review of locally and internationally available literature related to deflection data collection, interpretation and analysis was undertaken. This provided a clearer understanding of the research undertaken in the past and closed the existing knowledge gap on structural performance. Furthermore, an initial research study carried out to examine the structural performance of flexible pavements using TSD for network-level analysis, generated some interesting findings that deflection data collected by TSD can be easily transformed into FWD- equivalent Structural Numbers. Additionally, this study has successfully established a correlation between maximum deflections TSD-D0 and FWD-D0 derived from the two devices. In this study, TMR's tolerable-deflection FWD equations are transformed to TSD- based equations through the derived TSD and FWD relationship. Subsequently, a methodology has been developed to predict structural remaining life from the measured TSD deflection data. The major part of this research was devoted to develop structural deterioration models for flexible pavements for Queensland roads. The structural deterioration model is a critical input in the pavement management system for predicting structural performance and pavement rehabilitation needs. A simple, practically implementable concept was adopted in this study to develop the models. These deterioration models were developed by using five years of time series TSD deflection data as a function of other readily available data through adopting data mining techniques and tools. This study successfully established robust and simple structural deterioration models that will enable road agencies to incorporate available TSD data in the pavement management system for developing network-level pavement investment strategies. Furthermore, the deterioration models were developed as a function of maximum deflection (D0) and adjusted structural number. These models were useful as most of the currently available pavement management software can incorporate these models without making any major changes to the program. In conclusion, this is the first research carried out in Queensland using a large amount of network level data collected by TSD for developing structural deterioration models. The findings of the research make a significant and original contribution to knowledge and narrow the knowledge gap of practically using structural data at a network level. The structural deterioration models developed in this research will assist road agencies in the sustainable management of ageing road assets and provide them with greater insights into pavement conditions.