Prediction is always a challenging task; it gets harder especially in mining where lots of complexities and uncertainties are present in the system. Optimising the production output by adhering to the ore quality, minimising fuel consumption towards operational cost reduction, maximising utilisation and minimising the idle time of the fleets are a few major goals in the mining industry. However, all these things depend upon the optimal distribution of resources and equipment in appropriate places. Though manual allocation can be one solution, optimal result is not always achieved because it is difficult to optimise so many parameters on a day-to-day basis. The present research proposes a multistage and multi-objective optimisation approach based on mixed integer nonlinear programming to achieve the aforesaid goals. The experimental results show the efficacy of the method, and it is also implemented in one real mine scenario where all the above-mentioned goals are achieved.