A methodology to monitor the real-time cutting forces generated by a longwall shearer is developed. Longwall shearers are the single productive element in an underground longwall coal mine, and are a major source of unscheduled system downtime. The cutting forces constitute the interaction between the longwall and the coal face. Knowledge of the forces can improve the reliability and productivity of longwall mines by providing a means of shearer condition monitoring. A Kalman filter based cutting force estimator is developed, and then tested using a comprehensive shearer simulation incorporating a dynamic representation of the cutterhead. Subsequently, prior work on autonomous shearer guidance with measured pick forces is extended by employing the estimated cutting forces to track changes to the shearer's horizon. This is validated by simulation. The force estimation methodology uses an extended Kalman filter to blend noisy measurement information with a nonlinear system model, the latter combining the dynamics of the shearer and its inputs. A detailed model is developed of the shearer rigid body dynamics, and of the measurements required for observability. The formulation avoids the use of sensors at the coal cutting interface. The methodology uses proven, practicable and reliable sensor technologies, making it suitable for longterm force monitoring. A comprehensive simulation of a longwall shearer is developed using the MATLAB(R) Simulink(R) and MSC.visualNastran 4DTM software environments. This incorporates dynamic representations of the major structures and inputs, including the interaction between the cutterhead and the coal face. The simulation is used to investigate the cutting forces in a number of different operational scenarios. This facilitates the construction of a basic dynamic model of the forces, and provides insight into the relationship between the cutting forces and the condition of the cutting interface. The cutting force estimator is tested using measurements from the shearer simulations. It is shown that the cutting force estimates track the fiducial forces well in each scenario. A number of modifications to the force models are also shown to improve the performance of the estimator. Sensitivity analyses are conducted to investigate the effect of tuned process noise covariance and model parameter errors on the estimates. The application to the problems of shearer condition monitoring and horizon tracking are considered. The first is demonstrated by detecting the effects of a worn pick and a seam hardness gradient on the force estimates, and the second by using the force estimates to track the motion of the cutterhead relative to an embedded stone band. The results show that it is possible, in principle, to estimate the cutting forces generated by a longwall shearer from indirect measurements using a Kalman filter methodology. Furthermore, the force estimates reveal changes to the cutting environment, and can track the vertical motion of the shearer as it operates. This thesis represents the first step in the development of a shearer cutting force estimator. The ultimate goal of the research is to improve the reliability and productivity of Australian longwall operations, using the estimated forces to monitor shearer condition. Autonomous horizon tracking with estimated forces can lead to further improvements to longwall productivity, and to the health and safety of mine personnel. A number of recommendations are made for future field validation, and development of an automated shearer condition monitoring system based on the cutting force estimates.