A parameter correction method includes: obtaining, from a variability-aware simulation, a simulation result value of a predetermined product performance for a reference candidate value set concerning statistics of predetermined product characteristics; calculating a likelihood by substituting the reference candidate value set, the obtained simulation result value, statistics of measurement values of the predetermined product characteristics and a measurement value of the predetermined product performance into a likelihood function that is defined from a probability density function for the statistics of the predetermined product characteristics and a probability density function for the predetermined product performance, and is a function to calculate a combined likelihood of the statistics of the predetermined product characteristics and the predetermined product performance; and searching for a reference candidate value set in case where the calculated likelihood becomes maximum, by carrying out the obtaining and the calculating plural times while changing the reference candidate value set.