Molecular motors play key roles in organizing the interior of cells. An efficient motor in cargo transport would travel with a high speed and a minimal error in transport time (or distance) while consuming minimal amount of energy. The travel distance and its variance of motor are, however, physically constrained by energy consumption, the principle of which has recently been formulated into the thermodynamic uncertainty relation. Here, we reinterpret the uncertainty measure (Q) defined in the thermodynamic uncertainty relation such that a motor efficient in cargo transport is characterized with a small Q. Analyses on the motility data from several types of molecular motors show that Q is a nonmonotic function of ATP concentration and load (f ). For kinesin-1, Q is locally minimized at [ATP] ≈ 200 µM and f ≈ 4 pN. Remarkably, for the mutant with a longer neck-linker this local minimum vanishes, and the energetic cost to achieve the same precision as the wild-type increases significantly, which underscores the importance of molecular structure in transport properties. For the biological motors studied here, their value of Q is semi-optimized under the cellular condition ([ATP] ≈ 1 mM, f = 0 − 1 pN). We find that among the motors, kinesin-1 at single molecule level is the most efficient in cargo transport.Biological systems are in nonequilibrium steady states (NESS) in which the energy and material currents flow constantly in and out of the system. Subjected to incessant thermal and nonequilibrium fluctuations, cellular processes are inherently stochastic and error-prone. To minimize detrimentaion effects, cells are equipped with a plethora of energy-consuming error-correction mechanisms. Trade-off relations between the energetic cost and information processing are ubiquitous in cellular processes, and have been a recurring theme in biology for many decades [1? ? -4].A recent study by Barato and Seifert [5] has formulated a concise inequality known as the thermodynamic uncertainty relation, the trade-off between energy consumption and precision of an observable from dissipative processes in NESS. To be more explicit, the uncertainty measure Q, defined as a product between the energy consumption (heat dissipation, Q(t)) from an energy-driven process in the steady state and the squared relative error of an output observable X(t), 2 X (t) = δX 2 / X 2 , is constant. It has been further conjectured that for an arbitrary chemical network formulated by Markov jump processes Q cannot be smaller than 2k B T ,The measure Q quantifies the uncertainty of a dynamic process. The smaller the value of Q, the more regular and predictable is the trajectory generated from the process, rendering the output observable more precise. The proof and physical significance of this inequality have been discussed [5][6][7][8][9][10]. In the presence of large fluctuations inherent to cellular processes, harnessing energy into precise motion is critical for accuracy in cellular computation.The uncertainty measure Q can be used to assess the e...