The utilization of renewable energy resources (e.g., solar energy, geothermal energy, and biomass energy) or recovery of waste heat (e.g., industrial waste heat and engine waste heat) is an effective way to deal with current energy and environmental problems. Organic Rankine cycle (ORC) is a promising technology among multiple heat-to-power methods. Working fluid is a key component of an ORC to complete the heat-to-power process and the screening of the existing working fluids and the design of new working fluids have been always the focus of the ORC research. Computer-aided molecular design (CAMD)-based working fluid screening simultaneously with the process optimization is an essential and promising way of improve the ORC performance. In the CAMD process, the prediction accuracy of working fluid properties is significant for the accurate working fluid screening, working fluid design, and cycle performance evaluation. In this study, an artificial neural network (ANN) -based property prediction method is developed and the property correlations with improved accuracy for critical pressure (P-c) and normal boiling temperature (T-b) are achieved. Based on the proposed prediction methods and correlations, a systematic methodology is proposed for fast performance evaluation of ORC using existing or new fluids. The proposed property prediction method, property correlations and ORC performance evaluation method are validated by comparing with REFPROP 9.1 database and/or previous methods. The validation results show that the absolute average deviation (AAD) of existing working fluid thermophysical properties calculated using the proposed method is significantly reduced. The MD of the thermal efficiency of the cycle using existing working fluid calculated using the proposed method is 26%-81% lower than those results achieved using the previous methods. In addition, the calculation time using the proposed method for ORC performance evaluation is 0.28s which is several orders of magnitude lower than those of the previous methods. (C) 2019 Elsevier Ltd. All rights reserved.