The incidence of diseases that are caused by fungal infection is gradually increasing, together with antibiotic abuse and the number of patients with hypoimmunity. The many challenges in clinical anti-fungi treatment include serious adverse effects and drug resistance. The mitochondria of fungi have been found to be closely associated with pathopoiesia and drug resistance. Hence, we investigated patterns in Candida mitochondrial genes codon usage bias to provide new information to guide anti-fungal research. According to the nucleotide composition results, most mitochondrial genes of the analysed Candida tended to use A/T bases rather than G/C bases. The relative synonymous codon usage values demonstrated that UUA, AGU, CCU, GCU, UGA, AGA and GGU were the common preferential codons of mitochondrial genes in 12 Candida species. Codon adaptation index (CAI) analysis indicated that the ATP9 of Candida parapsilosis had the highest value, and the ND6 of C. auris had the lowest value. The CAI clearly correlated with the codon bias index, except in C. maltose and C. viswanathii, and was significantly positively correlated with the average GC content. Together, our results suggested that the codon usage pattern is affected by multiple factors, among which GC content is critical. Nucleotide composition, selection pressure and mutation pressure influence codon bias in Candida mitochondrial genes, with dominant status to mutation pressure. Codon usage bias analyses of Candida mitochondrial genes may provide new insight into its evolution.