This paper examines the stimulative effect of central bank forward guidance—the promise to keep future policy rates lower than its policy rule suggests—when the short-term nominal interest rate is stuck at its zero lower bound (ZLB).We utilize a standard New Keynesian model in which forward guidance enters our model as news shocks to the monetary policy rule. Three key findings emerge: (1) Forward guidance is more stimulative at the ZLB when households believe the economic recovery will be strong. When households expect a weak recovery or initially have low confidence in the economy, forward guidance is less stimulative because interest rates are already expected to remain low; (2) Longer forward guidance horizons do not cause the stimulative effect to explode or reverse, but rather spread the effect across the entire horizon; and (3) Failing to include a ZLB constraint causes the model to substantially overstate the stimulative effect of forward guidance. Given those findings, we use Blue Chip survey data to compare our model’s predictions of the stimulative effect of forward guidance to data before and after the Fed’s historic policy announcement on August 9, 2011. The results in that case study provide an explanation for the Forward Guidance Puzzle—the claim that New Keynesian models overestimate the effect of forward guidance.