We investigate the impact of news sentiment on the price dynamics of natural gas futures. We propose a Local News Sentiment Level model, based on the Local Level model of Durbin and Koopman [Time Series Analysis by State Space Methods, 2001], to construct a running series of news sentiment from irregularly observed news items’ sentiments. We construct several return and variation measures to proxy for the fine dynamics of the natural gas futures prices. We employ event studies, Granger causality tests and several state-of-the-art volatility models to assess the effect of news on the returns, price jumps and the volatility. We find significant relationships between news sentiment and the dynamic characteristics of natural gas futures prices. Our findings are, among others, that the arrival of news in non-trading periods causes overnight returns, that news sentiment is Granger caused by negative semi-volatility and that news sentiment is more sensitive to negative than to positive jumps. In addition to that we find strong evidence that news sentiment severely Granger causes price jumps and conclude that market participants trade futures as some function of aggregated news. We augment volatility models with news sentiment measures and conduct an out-of-sample volatility forecasting study. The first class of models is the generalized autoregressive conditional heteroskedasticity models the second class is the high-frequency-based volatility models of Shephard and Shephard [J. Appl. Econ., 2010, 25, 197–231] and Noureldin et al. [J. Appl. Econ., 2012, 27(6), 907–933]. We adapt both models to account for asymmetric volatility, leverage and time to maturity effects. By augmenting all models with news sentiment variables, we find that including news sentiment in volatility models significantly improves volatility forecasts.