In charged-particle therapy treatment planning, the patient is conventionally modeled as variable-density water, i.e. stopping effective density ρS, and the planar integrated dose distribution measured in water (PID) is applied for patient dose calculation based on path length scaling with the ρS. This approximation assures the range accuracy of charged-particle beams. However, it causes dose calculation errors due to water nonequivalence of body tissues in nuclear interactions originating from compositional differences. We had previously proposed and validated a PID correction method for the errors in carbon-ion radiotherapy. In the present study, we verify the PID correction method for helium-, oxygen-, and neon-ion beams. The one-to-one relationships between ρS and the nuclear effective density ρN of body tissues were constructed for helium-, carbon-, oxygen-, and neon-ion beams, and were used to correct the PIDs to account for the dose calculation errors in patient. The correction method was tested for non-water materials with un-scanned and scanned ion beams. In un-scanned beams penetrating the materials, the dose calculation errors of up to 5.9% were observed at the Bragg peak region, while they were reduced to ⩽0.9% by the PID correction method. In scanned beams penetrating olive oil, the dose calculation errors of up to 2.7% averaged over the spread-out Bragg peak were observed, while they were reduced to ⩽0.4% by the correction method. To investigate the influence of water nonequivalence of body tissues on tumor dose, we carried out a treatment planning study for prostate and uterine cases. The tumor over-doses of 0.9%, 1.8%, 2.0%, and 2.2% were observed in the uterine case for the helium-, carbon-, oxygen-, and neon-ion beams, respectively. These dose errors could be diminished by the PID correction method. The present results verify that the PID correction method is simple, practical, and accurate for treatment planning of these four ion species.