Purpose: Current trend in implant dentistry is changing from external connection to internal connection. To evaluate the splinting of external and internal connection implant on screw loosening, 2-units prosthesis was fabricated with BioPlant $System^(R)$ of external connection type and Lifecore STAGE-1 Single Stage Implant $System^(R)$ of internal connection type. Material and Method: Experimental group is classified into three groups. 1) $G_1-EE$: 2-units prosthesis was fabricated with two Bioplant $System^(R)$ of external connection type. 2) $G_1-EI$: 2-units prosthesis was fabricated with one BioPlant $System^(R)$ of external connection type and one Lifecore STAGE-1 Single Stage Implant $System^(R)$ of internal connection type. 3) $G_1-II$: 2-units prosthesis was fabricated with two Lifecore STAGE-1 Single Stage Implant $System^(R)$ of internal connection type. In fabricating 2-units prosthesis, two hexed abutments are recommended when two implants are installed parallel, otherwise one hexed abutment is used on major occlusal force area and one nonhexed abutment is used on the other area. Since it is rare to find two implants being parallel, it is hard to fabricate prosthesis with passive adaptation using two hexed abutments. It is much more difficult to acquire passive adaptation when using hex abutment compared to nonhex abutment. To evaluate the influence of hexed and nonhexed abutment on screw loosening, 2-units prosthesis was fabricated with hexed and nonhexed abutment. Experimental group is classified into three groups. 1) $G_2-HH$: 2-units prosthesis was fabricated with two hexed abutments. 2) $G_2-HN$: 2-units prosthesis was fabricated with one hexed abutment and one nonhexed abutment. 3) $G_2-NN$: 2-units prosthesis was fabricated with two nonhexed abutments. Result: The results of comparing the detorque value after loading on a each prosthesis periodically are as follows. 1. In splinting group of external and internal connection implant, $G_1-II$ group demonstrated the biggest detorque value, followed by $G_1-EI$ group and $G_1-EE$ group. 2. There is no notable significance between external connection implant of $G_1-EI$ group and $G_1-EE$ group and also no significance between internal connection implant of $G_1-EI$ group and $G_1-II$ group. 3. $G_2-HH$ group showed higher detorque value than $G_2-HN;and;G_2-NN$ group. From the results, we can concluded that using both external connection and internal connection implant together is clinically acceptable and in order to acquire a good passive adaptation in fabricating 2-units implant prosthesis we can use two nonhexed abutments.