Lake restoration usually focuses on reducing external nutrient sources. However, when sediments contain nutrients accumulated over multiple years, internal nutrient release can delay restoration progress. In lake restoration and management, it is important to understand the dynamic relationship between nutrient concentrations in a lake and internal and external nutrient sources. In this study, we quantified external nutrient inputs through measurements and compared them with internal sediment release from simulation using the PCLake+ model. Additionally, we evaluated alterations in the internal nutrient release, lake nutrient concentrations, and algae biomass (chlorophyll-a) within the lake following varying degrees of reduction in external nutrient loads. The results demonstrate that the PCLake+ effectively simulated the lake's nutrient concentration and algae biomass. Based on the PCLake+ estimates, internal nutrient loads accounted for 51 % of the total nitrogen (N) and 80 % of the total phosphorus (P) loadings in Lake Erhai in 2019. In 2020, the total contributions were 43 % for TN and 72 % for TP. We simulated four scenarios where external nutrient inputs were reduced to 25 %, 50 %, 75 %, and 99.99 % of their original levels. The 40-year simulation showed that the lake's ecological system initially exhibited a fast internal response but reached equilibrium after eight years. P concentrations took longer to reach equilibrium compared to N concentrations, probably due to the stronger binding characteristics of P. To meet the water quality target in the future, it is necessary to reduce external N and P inputs into Lake Erhai by at least 23 % and 15 %, respectively, under current conditions. Although reducing external nutrient loads can indirectly lower internal nutrient loads, water management should address both external and internal loads simultaneously, as internal release cannot be effectively reduced by external reductions alone. Additionally, the lake's internal release may continue for several years, even with reductions in external inputs.