Camera trap method has been developed for monitoring wildlife, however, most studies using camera trap depend on baited camera sites to attract target wildlife. This is likely to bias estimates of population structure. We evaluated the use of non-baited camera trap for the estimation of herd composition of sika deer (Cervus nippon). Camera trap showed a distinct seasonal pattern in sex ratios (males/100 female), which remained lowest between May and October but increased in November. Sex ratios were influenced by the number of observed males, because the ratios were positively correlated with the number of males but not females. The number of males increased in autumn during rutting season. Fawns/100 female ratios showed a distinct seasonal pattern. Highest and lowest fawns/100 female ratios were obtained in November and May, respectively. The decrease of fawns/100 female ratios in May comparing with that in November may be because of the overwinter mortality of fawns. Camera trap method is superior in term of continuously conduct in long-term, collect reasonable seasonal patterns, automatically record large numbers of sample sizes and useful in all weather conditions.