Increasing regulations and social expectations of mines to minimize environmental impacts whilst ensuring a safe working environment and maintaining profitability has led to a higher degree of control throughout the mining and processing cycle. In recent years there has been an increasing trend with regards to the use of paste fill. Paste fill is defined simply as mine tailings (typically an effective grain size of approximately 5 μm) mixed with some form of binder. The tailings used to manufacture paste fill are unclassified; that is they are not graded or sized in any form. This lack of grading results in greater variation of the tailings and consequently the material characteristics. Traditionally, additional cement has been added into the paste fill blend to compensate for the variation of the grain size distribution and improve stability of fill exposures. Cement is the most expensive component of paste and can constitute between 15%–20% of the total cost of mining. Therefore, any reduction in the use of excess cement will result in obvious economic benefits. Within this study, artificial neural networks (ANNs) were applied to the prediction of fill strengths, and were based on the input parameters of cement content, solids content, curing time and grain size distribution. Correlations between the predicted and achieved strength of the paste for both Cannington mine and paste fill worldwide using ANNs were excellent. The use of ANNs as part of an integrated planning tool for the design of backfills has also been discussed in the paper.