Rain-dependant farming is prevalent in semiarid tropics and in Southern parts of India; farmers widely cultivate sorghum without scientific understanding of distribution as well probable receipt of rainfall. There is lack of knowl-edge about correct 'sowing window' which often leads to low yield and complete crop failure. The objective of this paper was to assess the quantum of rainfall received during the last 20 years and determine the probable on set and continuity of rainfall on weekly basis so as to raise a crop under rain-dependant situations. The field experiment was laid out in Factorial Randomized Block Design (FRBD) with 3 replications and data were statisti-cally analysed in AGRES software. The treatments comprised of Sorghum (Sorghum bicolor L.) (C1) and little millet (Panicum sumatrense L.) (C2) as factor one (crops) and sowing window based on standard meteorological week (SMW) viz., sowing at 31st SMW based on farmer's practice (M1), Sowing at 33rd SMW based on 50% rainfall prob-ability with 100% recommended doses of fertilizers (M2), Sowing at 38th SMW based on 75% rainfall probability with 100% recommended doses of Nitrogen, Phosphorus and potassium (M3), Sowing window based on current weather forecast-35th SMW (M4), as factor two. The study indicated Sorghum (C1) sowing on 38th standard mete-orological week (M3) is ideal to get maximum yield under rainfed condition.