Accurately predicting coal price is crucial for market decision-making, resource allocation, and macroeconomic regulation. This chapter first presents the detailed literature review to get the mainstream methods of the general studies of coal prices and the forecasting of coal prices. The general studies about coal prices are discussed from energy price bubbles, supply and demand factors, alternative energy factors, macroeconomic environment, and energy policy, respectively. The forecasting of coal prices is mainly classified into the traditional statistical models and the hybrid forecasting models, where the latter is further divided into the simple machine learning methods, the hybrid methods integrating multiple parameter models with machine learning methods, and the decomposition-integration models. Then, the decomposition-integration approach is employed to study coal prices prediction, where the representative Chinese coal price index is taken as an example. We use the Empirical Mode Decomposition (EMD) and Variational Modal Decomposition (VMD) to divide the coal prices into the low frequency and high frequency components, which are fitted by ARIMA model and machine learning method, respectively. The empirical results show that EEMD-ARIMA-LSTM model performs better in terms of forecasting accuracy while CEEMDAN-ARIMA-LSTM works better from the point of stability, which suggests us to select a better model from two dimensions of precision and efficiency respectively.