Time delay (TD) procedure is an effective procedure in teaching various skills to children with developmental disabilities. Moreover, research has shown that it is used with high treatment integrity (HTI). However, there are several barriers which may prevent delivery instruction with HTI. Therefore, this study was designed to compare the effectiveness and efficiency of TD instruction with HTI and low treatment integrity (LTI) in teaching object naming to children with autism. LTI was defined as not delivering prompt in 30% of all teaching trials. An adapted alternating treatments design was used in the study. Three 5 to 6 year old male students participated in the study. Results showed that all three children learned their target behaviors on the criterion level with both conditions. Mixed findings were obtained for the efficiency of instruction. Based upon evaluation of the findings, recommendations for practitioners are provided and future research needs are discussed.