Background: As part of an ongoing clinical service programme for pre-school children with developmental delay in an Asian developing country, we analysed the effect of three assessment tests, that is, Bayley Scale of Infant Development-II, Leiter International Performance Scale-Revised and Wechsler Preschool and Primary Scale of Intelligence-Revised-Chinese, on the stability of intelligence quotient (IQ) of children from pre-school through early childhood. Methods: The participants were 313 Taiwanese pre-school children with uneven or delayed cognitive profile and they were followed through early childhood. IQ stability was explored by different tests and among children of different clinical diagnosis: 168 children with non-autistic intellectual disability, 73 children with autism spectrum disorder, 58 children with mixed receptive-expressive language disorder and 14 children of other heterogeneous diagnoses. Stability of scores was evaluated using the "r"-squared for Pearson's coefficients to see the correlation between initial IQ (IQ1) and follow-up IQ (IQ2). Multiple linear regressions were also applied to see whether IQ1 had predictive ability for IQ2 and test-test difference in the total 313 children and each diagnostic subgroup. Results: Results revealed that mean IQ1 was 65.8 [plus or minus] 15.4 while mean IQ2 was 73.2 [plus or minus] 17.9 for the total 313 children. The IQs were stable across an average follow-up duration of 38.6 [plus or minus] 22.1 month from pre-school into early childhood. Patterns of positive correlations between IQ1 and IQ2 were noted by all the tests (r-squared = 0.43-0.5, all P less than 0.001) and in the majority of diagnostic subgroups. Multiple regressions analysis also revealed that IQ1 could predict IQ2 significantly in all the tests (all P less than 0.001). Discussion: After careful choice of appropriate initial test, stability of IQ in children with developmental delay was noted from pre-school through early childhood. In addition, the translated version of cognitive assessment was valid for the required context of an Asian developing country. With the current emphasis on early identification and intervention for pre-school children with developmental delay, this information bears merit in clinical practice.