Property market forecasts play a crucial role in modern real estate valuation methodologies and, consequently, flawed forecasts can have adverse impacts on the accuracy of valuations. This thesis identifies property industry inconsistencies in the formulation and application of office rent forecasts adopted in discounted cash flow (DCF) studies used to assess the value of commercial properties and the viability of proposed projects. Existing research on commercial property cycles and office property market modelling is examined in order to identify the dominant market drivers adopted by researchers. Forecasting techniques are also explored towards specifying space and rent models for the Brisbane CBD office market using the perceived dominant drivers as explanatory variables. Surveys of property valuers and developers are undertaken to underpin the selection of these variables. The implications of varying rent forecasts applied in DCF based valuation assessments are tested through the use of a case study involving four Brisbane office buildings. Innovative research is conducted through adopting geographic information system supported land use and historical valuation studies to delineate market precincts within the Brisbane CBD. The rent model is then re-estimated using precinct based office rent data to allow the generation of forecasts for the individual precincts. Out-of-sample accuracy test results for the precinct forecasts are compared with the results produced by the model specified using whole-of-city data. The literature reviews, surveys and model testing determine a relatively consistent range of dominant explanatory variables applicable to office markets. The case study, in a local context, confirms that varying forecasts do have a significant impact on property valuations. Tests of the forecast results generated by the Brisbane CBD model provide some evidence that more plausible office rent forecasts stem from the use of market models as compared with solely applying professional judgement based forecasts. Subject to data availability limitations, the precinct based rent model is found to produce rent forecasts superior to those generated by the whole-of-city model. Finally, the thesis makes a range of industry recommendations towards enhancing forecasts and recommendations are also made for potential future research projects.