Soil acidity is one of the main limiting factors for coffee production in Peruvian rainforests. The objective of this study is to predict the spatial acidity variability for recommending site-specific liming and phosphorus fertilization treatments. We analyzed thirty-six edaphoclimatic variables, eight methods for estimating liming doses, and three geospatial variables from 552 soil samples in the Pichanaqui district of Peru. Multivariate statistics, nonparametric comparison, and geostatistical analysis with Ordinary Kriging interpolation were used for data analysis. The results showed low coffee yields (0.70 ± 0.16 t ha⁻¹) due to soil acidification. The interquartile ranges (IQR) were found to be 3.80–5.10 for pH, 0.21–0.87 cmol Kg⁻¹ for Al⁺³, and 2.55–6.53 mg Kg⁻¹ for available P, which are limiting soil conditions for coffee plantations. Moreover, pH, Al⁺³, Ca⁺², and organic matter (OM) were the variables with the highest accuracy and quality in the spatial prediction of soil acidity (R² between 0.77 and 0.85). The estimation method of liming requirements, MPM (integration of pH and organic material method), obtained the highest correlation with soil acidity-modulating variables and had a high spatial predictability (R² = 0.79), estimating doses between 1.50 and 3.01 t ha⁻¹ in soils with organic matter (OM) > 4.00%. The MAC (potential acidity method) method (R² = 0.59) estimated liming doses between 0.51 and 0.88 t ha⁻¹ in soils with OM < 4.00% and potential acidity greater than 0.71 cmol Kg⁻¹. Regarding phosphorus fertilization (DAP), the results showed high requirements (median = 137.21 kg ha⁻¹, IQR = 8.28 kg ha⁻¹), with high spatial predictability (R² = 0.74). However, coffee plantations on Ferralsols, with Paleogene parental material, mainly in dry forests, had the lowest predicted fertilization requirements (between 6.92 and 77.55 kg ha⁻¹ of DAP). This research shows a moderate spatial variation of acidity, the need to optimize phosphorus fertilization, and an optimal prediction of liming requirements using the MPM and MAC methods, which indicate high requirements in the southwest of the Pichanaqui district.