The absence of any further fertility decline in Kenya in recent past has alarmed the demographic community. While such phenomenon is not new, it is also possible that the observed fertility as measured by TFR may be due to flaws in the in distortion of TFR from changes in the timing of childbearing. On the other handWhile, there may have been a real reversal in fertility decline that could arise from change in fertility preferences. However, tracing fertility trends by traditional measures (such as TFR) in early stages of demographic transition is speculative and uncertain even if data is of good quality. This study uses birth history data from the 1998 and 2003 KDHS to examine trends in family building patterns. The main conclusion is that fertility rates increased among women in their middle age (25-34) for those in parities 4 and 5 but declined for both younger and older women. From a methodological perspectives, the use of parity progression ratios uncovers patterns that may be difficult with the use of age- specific fertility rates. Parity Parity-specific analysis is consistent with the sequential nature of childbearing and approximates the family-building behaviour of real cohorts.