Marine sediments contain considerable amounts and different types of magnetic mineral particles. Magnetic minerals in sediments may be statistically aligned to the direction of the ambient geomagnetic field so that sediments potentially preserve geomagnetic intensity records in the past. However, different types of magnetic minerals should preserve the remanent magnetization in different manners. And the compositional variation of magnetic mineral assemblages in marine sediments may hinder us from extracting reliable geomagnetic paleointensity records. We studied a sediment core taken from the Ontong-Java plateau, from which relative paleointensity (RPI) variations were estimated. The magnetic mineral assemblages of the sediment core are principally a two-component mixture of terrigenous and biogenic magnetite. So it provides an opportunity to assess the influence that compositional variations in marine sediments could bring to RPI estimations and thus to distinguish different contributions of the biogenic and terrigenous components to RPI recording in marine sediments. RPI obtained by normalizing natural remanent magnetization (NRM) with anhysteretic remanent magnetization (ARM) shows downcore decreases, and it has an inverse correlation with the ratio of ARM susceptibility (k_ARM) to saturation isothermal remanent magnetization (SIRM) (k_ARM/SIRM). This indicates that the RPI signal becomes apparently weaker with increasing proportion of biogenic magnetite. Moreover, NRM-ARM demagnetization diagrams show concave-down curvature, which indicates that the coercivity distributions of NRM and ARM are different. If we assume that the magnetization of the higher coercivity interval is mainly carried by biogenic magnetite while that of the lower coercivity interval is mainly carried by the terrigenous component, RPI recording efficiency of the biogenic component may be lower than that of the terrigenous component. The validity of this assumption was investigated by first-order reversal curve (FORC) measurements, transmission electron microscope (TEM) observations, low-temperature measurements, and extraction of silicate-hosted magnetic inclusion from the sediments, and the results proved that NRM of the higher coercivity interval is carried mainly by biogenic magnetite. But our conclusion contradicts with some previous studies using a similar method, which suggested higher RPI recording efficiency of the biogenic magnetic component than the terrigenous component [Ouyang et al., 2014; Chen et al., 2017]. Different concentrations of silicate-hosted magnetic inclusions due to different sedimentary environments might be a possible reason for the contradiction. The contribution of silicate-hosted magnetic inclusions to the magnetization is minor in our sediments (less than ~7% of SIRM). This contradiction remains to be studied further. Keywords: geomagnetic paleointensity, silicate-hosted magnetic mineral inclusion, biogenic magnetite, Ontong-Java plateau