Rapid and high-sensitivity detections of biological analytes are critically important to ensure timely diagnosis of disease and effective monitoring of public health. Although various new biosensing platforms have been established as alternatives to conventional laboratory methods, most of these biosensing platforms suffer from insufficient sensitivities that severely limit their wide applications. To improve the detection sensitivities of these biosensors, surface modifications based on poly(amidoamine) (PAMAM) dendrimers and rolling circle amplification (RCA) have been proven to be effective methods. In this thesis, surface modification strategies based on PAMAM dendrimers and RCA have been applied on three biosensing platforms, including enzyme-linked immunosorbent assay (ELISA), localized surface plasmon resonance (LSPR) sensor chip, and affinity membrane, to improve their detection sensitivities. For the ELISA platform, glass-bottom and poly(styrene) 96-well plates are surface modified by dendrimer-aptamer conjugates to improve detection performances of human platelet-derived growth factor-BB using ELISA. The results show that the ELISA performed using the modified 96-well plates presents a much broader linear detection range and a significantly lower limit of detection (LOD) than conventional ELISA plates. For the LSPR platform, the dendrimer and aptamer modification strategy is employed to surface modify LSPR sensor chips for sensitive detection of the SARS-CoV-2 virus, and an RCA-AuNPs complex is developed to amplify the detection signals. The results show that the modified chip can sensitively detect the SARS-CoV-2 virus with a LOD of 148 vp/mL, suggesting that the modified LSPR chip and signal amplification method can be used for early diagnosis of Covid-19. For the affinity membrane platform, nylon membranes with dendrimer and dual-RCA surface modifications are developed to detect Escherichia coli O157:H7 in food samples. The surface-modified membranes significantly reduce the detection time of the target bacteria to two hours instead of several days using traditional bacterial detection methods. In addition, the new membranes achieve higher sample throughputs (around 4-5 mL/s) with a lower LOD (10 cells/ 250 mL) in processing real-world food samples compared to other similar detection platforms. The excellent properties of our surface modification approaches may provide further advantages when employed in other platforms, such as target separation and enrichment, antifouling and antibacterial, and drug delivery applications.