In rice–rice cropping system, monitoring of soil fertility and nutrient use efficiency changes over time, has a key role for proper fertilizer recommendation and sustaining rice production. In order to investigate the effect of long term application of fertilizers on the fertility of paddy field, an experiment was conducted in a RCB design with 4 replicates in a lowland rice field of Guilan province for 15 years. A combination of 3 levels of N (0, 30, and 60 kg N ha-1 from Urea source), with 3 levels of P (0, 30, and 60 kg P2O5 ha-1 as triple super phosphate) and 2 levels of K (0, and 30 kg K2O ha-1 from potassium sulphate source) fertilizers have been broadcasted on paddled soils and mixed thoroughly with 5 cm of soil surface by hand. The main object of this study was to investigate the fertilizers effects on total N, available P and K and grain yield variation over 15 years. The results have indicated an increasing trend of N accumulation (even in control treatment) in the soil. Although application of 60 kg N ha-1 have increased soil nitrogen content over above mentioned period of experiment compared to the control treatment, but had no significant difference in comparison to 30 kg N ha-1. The amount of available potassium in all plots (even control treatment) has significantly increased over 15 years, while application of 30 kg K2O ha-1 did not significantly increase available K during the study compared to control treatment. This may be due to high K concentration in irrigation water and low rate of K application. Soil available phosphorus in control treatment have not shown meaningful changes over years, while application of 30 or 60 kg P2O5 ha-1 have increased available P proper to rate of P application. In other words, soil available P content has correlated with the rate of P application over years. Application of K fertilizer without N and P has tended to produce grain yield equal or less than control treatment, while application of nitrogen with K fertilizer, has increased grain yield compared to control or 30 kg K2O ha-1 treatments. Results have also shown that nitrogen use efficiency increase in 30 or 60 kg N ha-1 treatments, when accompanied with phosphorus and potassium fertilizers application. P use efficiency in 30 or 60 kg P2O5 ha-1 treatments did not show any notable value in the first few year of study, but gradually have increased as time had proceed. Key words: Rice long-term experiment, N, P, and K fertilizers, N use efficiency.