Objectives To assess the differences in development of fetal heart rate (FHR) parameters of FHR patterns (FHRP) A and B, in the second half of pregnancy and to evaluate the performance of the Sonicaid System 8002(R) with respect to the identification of these FHR patterns. Subjects and methods A total of 235 computerized FHR recordings were obtained in 29 uncomplicated pregnancies longitudinally studied from 24 weeks' gestation onwards. Nomograms were made of basal FHR, longterm (LTV) and short-term (STV) FHR variation during both FHRP A and B, using multilevel analysis. Results During gestation, basal FHR decreased in both FHRP A and B. LN and STV increased during FHRP B and slightly decreased during FHRP A near term. For both FHRP A and B, intrafetal variance was much smaller than the total variance, indicating intrafetal consistency. At all gestational ages, the Sonicaid System poorly identified A and B patterns. This was largely due to a strong correlation of both LTV and STV with FHR. Correction for rate resulted in 'correct' identification in more than 85% of recordings. Conclusions We provide reference ranges for FHR, LTV and STV during both FHRP A and B from 24 weeks' gestation onwards. During FHRP A, LTV and STV are in 50% of cases below the normal range of overall FHR variation, hampering adequate assessment of the fetal condition. Therefore, we stress the importance of including B patterns in the analysis. There is no physiological background for the assignment of periods of low and high variation, as is done by the Sonicaid System 8002(R). However, after 30 weeks' gestation, computerized identification of patterns A and B may be possible using techniques as presented in this article.