Drought stress, uncertain and variable rainfall, low soil quality and nutrient deficiencies are among principal constraints for enhancing and sustaining agronomic productivity in rainfed farming in semiarid tropical regions of India. Therefore, long-term (1985–2004) effects of cropping, fertilization, manuring (groundnut shells, GNS; farmyard manure, FYM) and integrated nutrient management practices were assessed on pod yields, nutrient status and balances for a groundnut (Arachis hypogaea) monocropping system. The five nutrient management treatments were: control (no fertilizer); 100 % recommended dose of fertilizer (RDF) (20:40:40 N, P, K); 50 % RDF + 4 Mg ha−1 GNS; 50 % RDF + 4 Mg ha−1 FYM and 100 % organic (5 Mg ha−1 FYM). All treatments were replicated four times. The experiment was conducted at Anantapur district, Andhra Pradesh on an Alfisol using a Randomized Complete Block design. The gap in pod yields between control and different nutrient treatments widened with increase in duration of cultivation. Use of diverse fertilizer and manurial treatments produced significantly higher yields than control (P < 0.05). Amount and distribution of rainfall during critical growth stages was more important to agronomic yield than total and seasonal rainfall. Thus, the amount of rainfall received during pegging stage (r = 0.47; P < 0.05) and pod formation stage (r = 0.50; P < 0.05) was significantly correlated with the mean pod yields. Whereas, use of diverse fertility management practices improved nutrient status in soil profiles (N, P, K, S, Ca, Mg, Zn, Fe, Mn and B) after 20 years of cropping, yet soil available N, K and B remained below the critical limits. Long-term cultivation also caused deficiency of S, Zn and B, which limited the groundnut productivity. Crop removal of N, P and K during 20 years of cultivation was more in 50 % RDF + 4 Mg ha−1 GNS at 523, 210 and 598 kg ha−1, respectively. With the exception of control, there was a positive nutrient balance of NPK in all other treatments. Higher positive balance of N and K were observed in 50 % RDF + 4 Mg ha−1 GNS (616 and 837 kg ha−1, respectively), and those of P in 100 % RDF (655 kg ha−1) treatment. There was also a net depletion of available S, Zn, Cu and Mn, but a buildup of available Ca, Mg and Fe. Application of equal amount of GNS was as effective as or even better than FYM in terms of pod yields and nutrient buildup in the soil.