We report results from a deep Jansky Very Large Array (JVLA) search for 12CO J = 1–0 line emission from galaxies in a candidate galaxy cluster at z ∼ 1.55 in the COSMOS field. We target four galaxies with optical spectroscopic redshifts in the range z = 1.47–1.59, consistent with the likely redshift for the candidate galaxy cluster. Two of these four galaxies, ID 51613 and ID 51813, are nominally detected in CO 1–0 line emission at the 3–4σ level. We find CO luminosities of (2.42 ± 0.58) × 1010 K km s−1 pc2 and (1.26 ± 0.38) × 1010 K km s−1 pc2, respectively. Taking advantage from the clustering and expanded 2-GHz bandwidth of the JVLA, we perform a search for emission lines in the proximity of optical sources within the field of view of our observations (60 arcsec). We limit our search to galaxies with KS < 23.5 (AB) and zphot = 1.2–1.8. We find two bright optical galaxies, ID 51207 and ID 51380, to be associated with significant emission line peaks (>4σ) in the data cube, which we identify with the CO 1–0 line emission. To test the reliability of the line peaks found, we performed a parallel search for line peaks using a Bayesian inference method. Both CO line emitting candidates are identified with probabilities of 13 per cent and 72 per cent that there are line emitting sources in each case, respectively. Monte Carlo simulations show that such associations are statistically significant, with probabilities of chance association of 3.5 per cent and 10.7 per cent for ID 51207 and ID 51380, respectively. Modelling of their optical/IR spectral energy distributions indicates that the CO detected galaxies and candidates have stellar masses and star formation rates in the range (0.3–1.1) × 1011 M⊙ and 60–160 M⊙ yr−1, with star formation efficiencies comparable to that found in other star-forming galaxies at similar redshifts. By comparing the space density of CO emitters derived from our observations with the space density derived from previous CO detections at z ∼ 1.5, and with semi-analytic predictions for the CO luminosity function, we suggest that the latter tend to underestimate the number of CO galaxies detected at high redshift. Finally, we argue about the benefits of future searches for molecular gas line emission in clustered fields with upcoming submillimetre/radio facilities.