Objective: To investigate the effectiveness of the strengths and difficulties questionnaire (SDQ) in the screening for autism spectrum disorders (ASD) and to find optimal scoring methods. Methods: The SDQ were given to parents of 4-12-year-old children with ASD (n = 67), internalizing (n = 49), externalizing disorders (n = 87), and to a community sample (n = 114). The total difficulties score and five subscales scores were compared between the four groups. Two additional scoring methods were calculated: prosocial behavior score minus peer relation problems score and sum of prosocial behavior score and emotional symptoms score. Receiving operating characteristics (ROC) analysis was used to find an effective way of screening for ASD. Results: All five subscales of the SDQ differed significantly between the ASD and the community samples. Only prosocial behavior and emotional symptoms subscales scores in the ASD group differed significantly from both internalizing and externalizing disorders groups (p < 0.001). The total difficulties score reliably differentiated the ASD group from the community sample, but not from children with psychiatric disorders. Additional scoring methods' problems showed the best area under the curve (AUC) in ROC analysis. Conclusion: The SDQ discriminates those with ASD well from other children with and without psychiatric disorders. Applying of additional algorithms is more effective rather than using the total difficulties or subscales scores alone.