Four studies indicated that moods and self-esteem can influence counterfactual thoughts. This was shown for counterfactuals generated for hypothetical situations (Study 1), for recalled life events (Study 2), and for agreement with counterfactual statements after laboratory tasks (Studies 3 and 4). High self-esteem (HSE) and low self-esteem (LSE) persons generated (Studies 1 and 2) or agreed to (Studies 3 and 4) more downward (worse than actuality) than upward (better than actuality) counterfactuals when in good moods, but they diverged in reactions to bad moods: HSE persons thought more about downward counterfactuals, whereas LSE persons thought more about upward counterfactuals. HSE persons felt better after generating downward counterfactuals (Study 2) and took longer to agree to analogous statements (Studies 3 and 4) in bad moods, suggesting attempts at mood repair.