The correlation between lactation and test-day(TD) milk yields is high (El Faro and de Albuquerque 2003; Ilatsia et al. 2006). Some studies have shown minimal change in ranking of sires and cows (Swalve 1995; Meyer et al. 1989; Kaya et al. 2003). In developing countries, there is limited level of milk recording, and use of TD models would result in reduced cost of recording as we could have longer intervals between milk recording and less frequent collection of milk samples. In this way, the amount of information that can accrue from incorporating the majority of smallholders who have small herd sizes would be large. In Kenya, official milk recording schemes in cattle for smallholders is non-existent. A majority of dairy farmers are smallholder producing some 56% of total milk and 80% of the total marketed milk nationally (Omore et al. 1999) based on small herd sizes of two to three animals in about 1 ha land size. The dairy industry is a significant source of employment in Kenya. With a suitable policy framework, for example, subsidies and economic incentives coupled with suitable TD models, the level of milk recorded will increase, which is crucial for any meaningful accurate genetic evaluation. Increased genetic gain and improved profits can accrue from using TDMY on dairy heifer at early stage of lactation, unproductive cows would be culled early, and there will be decreased generation interval. In Kenya, we are yet to embrace the use of TD observations instead of aggregated 305-day production records despite several studies having shown advantages (Ilatsia et al. 2006; Mostert et al. 2006). If adopted, the country stand to gain in that ranking of animals could change significantly as observed in other countries (Schaeffer et al. 2000). Breeding programs in Kenya are based primarily on milk production, and therefore, accurate measurement or prediction of milk yield is essential for proving bull faster and eventually high genetic gain. Dairy yield prediction is a current challenge, which has been improved using different statistical methods. Recently, artificial neural networks (ANNs) have been employed as an alternative method of milk yield prediction.