Most Indian soils are categorized medium to high in K status, response of rice to applied K varies depending on the K status of the soil. The traditional practice is to apply the full dose of K basally for the rice crop, but it may not be available to plants in required quantity at the critical growth stages, since competition between micro-organisms and crop plants, luxury consumption, leaching losses and fixation process reduce the availability of K. Considering these points, to increase its efficiency, field experiments were conducted in different agro-climatic regions of India to study the effect of split application of K on rice. Results indicated that in Tamil Nadu and Orissa, application of K either in two splits (basal and panicle initiation) or three splits (basal, active tillering and panicle initiation) recorded higher grain yield. Application of K in two splits (basal and active tillering) in Kerala, Punjab and Uttar Pradesh, and in Himachal Pradesh (Active tillering and panicle initiation) recorded the highest grain yield. But in West Bengal and Madhya Pradesh, application of K in 3 splits (basal, active tillering, pre-flowering/panicle initiation) gave the best results. Depending on the initial K status and soil type, K has to be applied in two or three splits to increase its use efficiency and grain yield of rite. Low N use efficiency is common in lowland rice soils. This opens up the possibility of formulating new N:K granules for application to lowland rice.